This month, two experimental papers, in which I also contributed, have been published:

• Soogil Lee, Min-Gu Kang, Dongwook Go, Dohyoung Kim, Jun-Ho Kang, Taekhyeon Lee, Geun-Hee Lee, Nyun Jong Lee, Sanghoon Kim, Kab-Jin Kim, Kyung-Jin Lee, Byong-Guk Park
  Efficient conversion of orbital Hall current to spin current for spin-orbit torque switching
  Communications Physics 4,234 (2021)

• Dongjoon Lee, Dongwook Go, Hyeon-Jong Park, Wonmin Jeong, Hye-Won Ko, Deokhyun Yun, Daegeun Jo, Soogil Lee, Gyungchoon Go, Jung Hyun Oh, Kab-Jin Kim, Byong-Guk Park, Byoung-Chul Min, Hyun Cheol Koo, Hyun-Woo Lee, OukJae Lee, Kyung-Jin Lee
  Orbital torque in magnetic bilayers
  Nature Communications 12, 6710 (2021)

These papers propose ways to experimentally measure the orbital current contribution to the torque on the magnetization in nonmagnet/ferromagnet heterostructures: Cr/(Pt)/FM in the first study and (Ta,Pt)/FM in the second study, respectively. The main idea is if there is an orbital Hall effect in the nonmagnetic layer, the measured torque will substantially deviate from the theoretical prediction based solely on the spin Hall effect contribution. This can lead to gigantic enhancement (first paper) or a sign change (second paper) of the torque efficiency compared to a case when there is only a spin-injection contribution to the torque.

Both experiemnts have been performed in the beginning of 2019 and the key reuslts were obtained during the summer in 2019. Around that time, the idea of “orbital torque” was still very new idea for us and we did not have good intuition on how the orbital torque would behave and how large it would be. It is because the only theoretical study available was a case study on a tight-binding model that I have used in the original orbital torque paper.

Fortunately, that was the time when I came to Forschungszentrum Jülich as a postdoc in order to develop a code and perform first-principles calculations of the orbital torque. Ideally, I wanted to investigate Cr or Ta, which were used in the experiments, but I did not have good experience in the simulation of real materials yet, and I was not sure which structure I should assume (it wasn’t clear from the experiment either). Instead, when I just arrived at Jülich, I heard from my colleagues that they were already investigating Fe/W(110) because of on-going collaborations with STM and ARPES group. Thus, I decided to take an W(110) film as a “prototypical example”. It was a good choice because W has been predicted to have the opposite signs for the orbital and spin Hall effects, which is exactly what we wanted. Detailed computational study took more than I expected but we ended up publishing a very good and well-explained new method + new result theory paper (at least in my opinion). In this study, we have formulated a general theory of angular momentum transfer when a spin-orbit coupled system is driven out of equilibrium and reach a steady state by an external electric field and have shown from first-principles implementations that a competition bewteen the spin injection and orbital-to-spin conversion in the ferromagnet can compete each other in the dynamics of the magnetic order parameter. I was very glad that the speculation that I have proposed in the preivious was demonstrated in real materials.

However, in the meantime, I and the experimental team thought that it would take too long if we wait for the theoretical support and decided to submit the experimental paper although a theory backup is not complete. That is why in the second paper, a theoretical analysis presented in the Supplemental Material does not use an up-to-date method that has been developed during 2019-2020. It rather presents “the best thing that we we could do at that time”, which is not exact but at the same time not too wrong.

On the other hand, the experimentalists who led the first paper decided to spend more time in order to design more reliable experiments to draw a conclusion with less ambiguity. Since a similar study was already submitted (the second paper), we had to make one/two step further in order to claim the originality of the work. As usual, it took quite longer than we planned, but fortunately that gave me more time to perform theoretical calculation. This situation made one-to-one comparisions between the theory and experiment possible. During this process, I truely enjoyed listening to opinions from the experimentalists and designing a logic that can verify their hypotheses from my theoretical calculations (computer simulation is at the end a kind of “virtual” experiment, so designing a logic is very important), which is otherwise not so easy to check in experiments alone.

On the experimental side, we also tried to measure the magnetic torque in various different ways. Although the quantiative results can be slightly different depending on the method, we checked that the main conclusion can be generally drawn independently of the measurement method. Also, many possibilities that can lead to an artifact have been rasied, which we checked thoroughly to make sure.

As a result, although the initial data was obtained in the summer of 2019, it took us more than 2 years to have the work published. It was because these papers report the orbital Hall torque in transition metal systems for the first time, and this led to lengthy peer review processes. Most of the criticisms have been very helpful in the sense that they made us look at the data in a more objective point of view (which is indeed the most crucial role of the peer review in science). At the end, this made our paper more convincing and well-recognized in the community. I believe that it is just our first step at the entrance of the area “orbitronics”. I don’t know exactly what we will encounter in the future (though I have my own speculations). I would be happy to see the progress of the field and catch the holy grail of the orbital current in the near future.