Gate controlled valley polarizer in bilayer graphene

Sign reversal of Berry curvature across two oppositely gated regions in bilayer graphene can give rise to counter-propagating 1D channels with opposite valley indices. Considering spin and sub-lattice degeneracy, there are four quantized conduction channels in each direction. Previous experimental work on gate-controlled valley polarizer achieved good contrast only in the presence of an external magnetic field. Yet, with increasing magnetic field the ungated regions of bilayer graphene will transit into the quantum Hall regime, limiting the applications of valley-polarized electrons. Here we present improved performance of a gate-controlled valley polarizer through optimized device geometry and stacking method. Electrical measurements show up to two orders of magnitude difference in conductance between the valley-polarized state and gapped states. The valley-polarized state displays conductance of nearly 4e2/h and produces contrast in a subsequent valley analyzer configuration. These results pave the way to further experiments on valley-polarized electrons in zero magnetic field.

In the manuscript, the authors experimentally investigated a purely gate-controlled valleypolarized 1D-channels in bilayer graphene. Surpringsly, the authors highly improved the performance of gated bilayer device by employing optimized geometric dimensions and sample stacking schemes, where the imperfections from device asymmetries and asymmetric gating are notable decreased.Thus, even without the application of external magnetic field, the channel resistance reaches 5.64 kΩhm, which is quite close to the expected quantized value (h/4e2≅6.45kΩhm).
The device fabrication is impressive and this work deserves to be published in Nature Communication. However, there is a misleading point that should be corrected by the authors. The effect of a perpendicular magnetic field in gated bilayer graphene in the Nature Nanotechnolgy paper is to effectively suppress the backscattering process due to the imperfection from various sources. Quantum Hall effect has not been involved in the whole process. A systematic theoretical understanding can be found in Frontiers of Physics 14, 23501 (2019), which needs be properly referenced in this paper by the authors. Additional comments are provided in below. 1. One of the coauthors, Jiabin Qiao, has been acknowledged, which is unreasonable. 2. In fact, some references are not related to this work. But some very related progresses including both experimental and theoretical works are missing, e.g. The manuscript reports experiments with valley-polarized electrons in bilayer graphene. It has been theoretically predicted more than ten years ago that if voltages of the opposite signs are applied to the graphene layers, at the edge between these regions topological states appeat. These topological states have been previously experimentally observed by other groups. However, at low magnetic field the transport was not ballistic, and the observations were performed in high magnetic fields, close or in to the quantum Hall effect regime. The current manuscript presents improvements which enabled observation of topological channels without a need to go to high magnetic fields.
The information on the journal reads: Nature Communications is an open access journal that publishes high-quality research from all areas of the natural sciences. Papers published by the journal represent important advances of significance to specialists within each field. The manuscripts presents high-quality research. Furthermore, topological matter is at the focus of attention of condensed matter physics. Therefore the manuscript is, in principle, eligible for publication in Nature Communications.
However, I do not find that it is currently well-written. Currently, it reads an an incremental extension of the previous works. Instead, the authors should build up an independent story (obviously still citing relevant articles). The argument of the authors is that since they see much higher current if the voltages at the two splits have opposite signs that if they have the same sign, then they must be observing the topological states. The observation is consistent with 4 states carrying one conductance quantum each. Both observations are consistent with the theory. I understand that it is difficult to argue from the experimental data that the states are of topological nature. However, are these conclusions supported by more observations? For example, is  Fig. 3? The authors need to improve their argumentation to build up the credible story. Significant improvements are needed before publication can be recommended.
Reviewer #3 (Remarks to the Author): Chen et al. reported electron transport measurement of electric field induced domain wall in ABstacked bilayer graphene. Quantized conductance was observed for the odd geometry of adjacent split gate regions. Furthermore, signatures of valley-valve were observed in the quadrapole-like gate configurations. Controlling the valley degree of freedom is of significant interest for valleytronics. Quantum valley Hall effect has been observed in both AB-BA domain walls and electrical field induced domain walls in bilayer graphene. Previous experiments on the latter one were elusive as no quantized conductance was observed at zero magnetic field. A rather high magnetic field has to be applied to make the 1D transport quantized. However, the big magnetic field induces concerns about the interpretation of the valley Hall physics, due to the rendering of regular edge states of standard quantum Hall effect. The current manuscript represents a big improvement in the physics of electrical field induced domain wall as no magnetic field is needed at all! The authors carefully examined possible origins of charge inhomogeneities and successfully addressed some of them. And as a result, much more convincing experimental data was obtained than those observed in previous device structures. What's even more exciting is that the quantized conductance of domain wall corresponds to a mean free path of ~700 nm. I would be very interested in learning the true mean-free-path if the channel length is not a limitation. However, I don't think the authors need to provide such data in this manuscript. A lot more efforts should be devoted to this direction, in my opinion.
In sum, I would recommend the publication of this manuscript at Nature Comm.

Reviewer #1 (Remarks to the Author):
In the manuscript, the authors experimentally investigated a purely gatecontrolled valley-polarized 1D-channels in bilayer graphene. Surprisingly, the authors highly improved the performance of gated bilayer device by employing optimized geometric dimensions and sample stacking schemes, where the imperfections from device asymmetries and asymmetric gating are notable decreased. Thus, even without the application of external magnetic field, the channel resistance reaches 5.64 kΩ, which is quite close to the expected quantized value (h/4e2≅6.45kΩhm).
The device fabrication is impressive and this work deserves to be published in Nature Communication.
We thank the reviewer for such encouraging and positive feedback.
However, there is a misleading point that should be corrected by the authors. The effect of a perpendicular magnetic field in gated bilayer graphene in the Nature Nanotechnology paper is to effectively suppress the backscattering process due to the imperfection from various sources. Quantum Hall effect has not been involved in the whole process. A systematic theoretical understanding can be found in Frontiers of Physics 14, 23501 (2019), which needs be properly referenced in this paper by the authors.
We thank the author for highlighting the specific reference which has now been included. The Frontier of Physics paper has been properly referenced (please see reply to additional comments).
We also agree that the reference to QHE has been somewhat confusing. The Additional comments are provided in below.
1. One of the coauthors, Jiabin Qiao, has been acknowledged, which is unreasonable.
Thanks for highlighting this. Indeed, there has been a mistake. This has been modified accordingly. We thank the reviewer for his support. We believe that we have addressed all issues and thank the reviewer for his support.
Reviewer #2 (Remarks to the Author): The manuscript reports experiments with valley-polarized electrons in bilayer graphene. It has been theoretically predicted more than ten years ago that if voltages of the opposite signs are applied to the graphene layers, at the edge between these regions topological states appear. These topological states have been previously experimentally observed by other groups. However, at low magnetic field the transport was not ballistic, and the observations were performed in high magnetic fields, close or in to the quantum Hall effect regime.
The current manuscript presents improvements which enabled observation of topological channels without a need to go to high magnetic fields.
The information on the journal reads: Nature Communications is an open access journal that publishes high-quality research from all areas of the natural sciences.
Papers published by the journal represent important advances of significance to specialists within each field. The manuscript presents high-quality research.
Furthermore, topological matter is at the focus of attention of condensed matter physics. Therefore, the manuscript is, in principle, eligible for publication in Nature Communications.
We thank the reviewer for this positive assessment of our work.
However, I do not find that it is currently well-written. Currently, it reads as an incremental extension of the previous works. Instead, the authors should build up an independent story (obviously still citing relevant articles).
As the reviewer states, our experiments constitute a major improvement to published experimental data.
The qualitatively new result is demonstration of chiral states at B = 0T. As stated in the manuscript, this will allow experiments without having the bilayer leads in the QHE and dominated by edge state transport.
We explain this as a result of improved device fabrication schemes together with a quantitative estimation of the effects of sample imperfections.
In order to further strengthen the paper, we have included now additional measurements with the device in valley analyzer configuration (see below).
The argument of the authors is that since they see much higher current if the voltages at the two splits have opposite signs that if they have the same sign, then they must be observing the topological states. The observation is consistent with 4 states carrying one conductance quantum each. Both observations are consistent with the theory. I understand that it is difficult to argue from the experimental data that the states are of topological nature.
However, are these conclusions supported by more observations?
Real proof for topological states would be the demonstration of a valley analyzer.
The experimental challenge hereby is that two high quality junctions are required.
In order to further strengthen the paper, we decided to show additional data demonstrating the chiral character of the states in fig 5 (and data in SI). Please see Please note that the reason why we do not combine two diagonal gates in 2nd pair to measure the above two state in one measurement is that charge neutrality points in upper and lower parts are not always the same. This may also be the cause why we do not achieve so highly resistances when the 2 nd gate pair is in evenconfiguration (insulating state). Please see Supplementary Information part 8 for more details. We thank the reviewer for highlighting this point which was not explained in the manuscript. We estimate induced charge carrier density and displacement field from the capacitive coupling as follows: The relations can be derived: resulting in a saturation of resistance. Importantly, we show here the raw data without any correction to series resistance caused by contact resistance and resistance of bilayer graphene.
As described in the main text, we can estimate the series resistance as the measured value at high density. Subtracting this value from the raw data yields a number very close to the expected quantum resistance. Fig. 4 not the mirror image of Fig. 3?

Why is
The reviewer hints on an important point, which we hope to make clearer in the updated version of the manuscript. The authors need to improve their argumentation to build up the credible story.
Significant improvements are needed before publication can be recommended.
We added two sets of important data which highlight the performance of our devices and demonstrate the chiral character of 1D-states: -In SI we include now additional 'mirror-image' results as proof of gate controllable states.
-Moreover, direct evidence for chiral states are added by measurement data with the device in valley-analyzer configuration.
We are confident that the additional data strengthens the paper significantly, and sincerely hope that the reviewer finds this convincing as well. We sincerely thank the reviewer for this very positive feedback.
What's even more exciting is that the quantized conductance of domain wall corresponds to a mean free path of ~700 nm. I would be very interested in learning the true mean-free-path if the channel length is not a limitation.
We should recall that the gate length in our device is ~400nm. We have estimated the mean free path from analysis of standard transport measurements, i.e. without operating the local gates.
We agree that it would be interesting to measure a series of devices with increasing gate length to observe a crossover from ballistic transport to backscattered dominated devices. Yet, this would imply a tremendous amount of additional work which we hope to avoid.
However, I don't think the authors need to provide such data in this manuscript. A lot more efforts should be devoted to this direction, in my opinion.
We agree with the reviewer that further characterization of device parameters and performance is required to enable further experiments and to strive towards applications. Yet, this is beyond the scope of this paper.
In sum, I would recommend the publication of this manuscript at Nature Comm.
We thank the reviewer for this very positive statement.
Manuscript change list: 1. At page 2, in Abstract part, original sentence 'It has been predicted that signal reversal of…' is modified to 'Signal reversal of…'.
Also, the reference to QHE is revised as: 11. At page 10, the subtracted channel resistance is slightly modified from 5.64 kOhm to 5.72 kOhm after slight correction.
12. From page 10, in the first and second paragraph in Discussion part, some changes have been made and marked with red color and highlighted in yellow.
13. At page 11, discussion on p-n junction effect on the BLG resistance have been modified. Three previous references are deleted and replaced with new ones.