Abstract
Topological materials and their unusual transport properties are now at the focus of modern experimental and theoretical research. Their topological properties arise from the bandstructure determined by the atomic composition of a material and as such are difficult to tune and naturally restricted to ≤3 dimensions. Here we demonstrate that nterminal Josephson junctions with conventional superconductors may provide novel realizations of topology in n−1 dimensions, which have similarities, but also marked differences with existing 2D or 3D topological materials. For n≥4, the Andreev subgap spectrum of the junction can accommodate Weyl singularities in the space of the n−1 independent superconducting phases, which play the role of bandstructure quasimomenta. The presence of these Weyl singularities enables topological transitions that are manifested experimentally as changes of the quantized transconductance between two voltagebiased leads, the quantization unit being 4e^{2}/h, where e is the electric charge and h is the Planck constant.
Similar content being viewed by others
Introduction
Josephson junctions created by coupling two superconductors through a weak link have been studied extensively for many years^{1,2,3}. The current across a Josephson junction yields information about the Andreev bound states (ABS) forming at the junction^{4,5,6,7}. In turn, the ABS spectrum is determined by the properties of the junction and the superconducting leads. For instance, if the leads are topologically nontrivial, the Josephson effect may be used to probe these topological properties. In particular, the 4πperiodicity of the supercurrent indicates the presence of topologically protected zeroenergy Majorana states^{8,9,10,11}, which may arise in onedimensional spinless pwave superconductors, semiconductor nanowires with proximityinduced superconductivity or at the surface of bulk materials.
The rapidly growing field of threedimensional (3D) Weyl semimetals deals with a bandstructure that exhibits conical energy gap closings: Weyl points^{12,13,14,15,16}. Unlike the Dirac point in graphene^{17} that may be gapped out through an appropriate coupling, isolated Weyl points are topologically protected. They can be regarded as monopoles with a positive or negative charge. A topological invariant—the Chern number—defined on a surface in momentum space characterizes the total charge of the monopoles it encloses. These monopoles give rise to many unusual features, such as chiral edge states and associated surface Fermi arcs^{14}. Topological protection guarantees that the only way to induce a gap is either to annihilate two Weyl points of opposite charge by bringing them together, or to couple two cones at a finite distance in momentum space through breaking of momentum conservation^{16}.
In this paper, we show that multiterminal Josephson junctions may be topologically nontrivial even if the superconducting leads are topologically trivial and no exotic materials are used to make the junction. Thus, the junction itself may be regarded as an artificial topological material, which displays Weyl singularities, when the energy of the lowest ABS goes to zero at certain values of the superconducting phases such that the gap in the spectrum closes. Below, we also show that their topological property can be easily probed by the transconductance between two voltagebiased leads, which is proportional to the Chern number.
Results
The topology of the bound state spectrum
We consider a junction with n superconducting leads connected through a scattering region (Fig. 1a). The leads α={0, 1, …, n−1} have the same gap Δ, though they may differ in the phase of the superconducting order parameter, ϕ_{α}. Due to gauge invariance, only n−1 phases are independent, hence we may set ϕ_{0}=0. Likewise we choose to focus on a short scattering region, characterized by an energyindependent scattering matrix in the basis of N=∑_{α}N_{α} transport channels, N_{α} being the number of channels in contact α. In a longer scattering region described by an energydependent , more bound states would appear at finite energy. However, the presence of Weyl singularities at zero energy only depends on . We also assume timereversal symmetry, such that , as well as spinrotation symmetry. Note, although, that our predictions are robust even if those symmetries are broken, see Discussion section.
The junction hosts a set of spindegenerate ABS, indexed by k, whose energies E_{k}≥0 are determined from the equation^{4}
where χ=arccos(E/Δ), and is a diagonal matrix that assigns to each channel the phase factor of the corresponding terminal. The spindegenerate ABS energies (Fig. 1b) are periodic in all phases with a period 2π. The total phasedependent energy of the junction reads E=∑_{kσ}(n_{kσ}−1/2)E_{k}, n_{kσ}=0, 1 being the occupation of the state k with spin σ.
Zeroenergy states are most easily described by making use of the mapping from states with spin σ at energy E≥0 to states with spin −σ at energy −E. Then a gap closing at a certain value of the phases, , corresponds to the crossing of two (singly degenerate) states. Thus, the zeroenergy state is doubly (spin) degenerate. We can then describe the lowest ABS band in the vicinity of the gap closing by the twobytwo Weyl Hamiltonian (Supplementary Note 1),
with the Pauli matrices in the basis of the two degenerate states corresponding to eigenvalue E=0. The fields h_{i} depend linearly on through the real matrix .
The form of the Weyl Hamiltonian in equation (2) indicates that we need at least three parameters to tune the system to the degeneracy point, h_{i}=0. Thus, for four terminals with three independent phases, the Weyl singularities appear as points in the 3D phase space. For five terminals, the Weyl singularities occur in general as onedimensional curves in the fourdimensional (4D) space of phases. This opens up the possibility to realize a topological material in arbitrary dimensions. Note that such multiterminal Josephson junctions cannot be characterized by means of the standard periodic table of topological semimetals^{18,19}, due to the distinct behaviour of the quasimomenta under particle–hole symmetry. A proper classification could be envisioned along the lines of ref. 20.
The topology of the junction can be characterized by a set of Chern numbers. A Chern number may be defined in the twodimensional (2D) subspace of two phases ϕ_{α} and ϕ_{β}, through the local Berry curvature of the ABS. The Berry curvature for the bound state k with spin σ is related to its wave function through
Note that does not depend on spin (Supplementary Note 2). The total Berry curvature of the manybody superconducting state then reads , in analogy to the expression for the energy. For fixed occupations n_{kσ}=0, 1, the integral of the Berry curvature over the elementary cell yields an integer, the Chern number
with .
Since the Weyl singularities appear as points in the 3D phase space, a third phase ϕ_{γ} may be used to tune the system through Weyl points, thus changing the Chern number. We see that a given band k contributes to the total Chern number with when it is empty, and with when it is doubly occupied, whereas it gives a zero contribution if there is a single quasiparticle in the band.
Probing Weyl points through quantized transconductance
Importantly, the current response of the junction with slowly varying phases reveals the Chern number. Biasing lead β with a voltage gives rise to the instantaneous current to contact α (Supplementary Note 2 and Supplementary Fig. 1)
where . The first term corresponds to the adiabatic current and the second term is the first order correction in the phase velocity. Let us now apply constant voltages to two leads. For incommensurate voltages, the two phases uniformly sweep the elementary cell. In the d.c. limit, the adiabatic current contribution then averages out, and the Berry curvature is replaced by its average value. Thus, we find that the d.c. current is linear in the voltages, and the transconductance is defined by the Chern number
Equation (6) shows that multiterminal junctions exhibit a d.c. current response typical for the quantum Hall effect, although based on different physics. The transconductance quantum is four times bigger than in the quantum Hall effect, which can be traced to the 2e charge of the superconducting Cooper pairs and the presence of two spin bands. To extract the small d.c. signal, the averaging time needs to be sufficiently long. The relevant time scale is determined by the lowfrequency current noise (see Supplementary Note 3).
Generally, at low temperatures one expects relaxation processes to bring the system to the ground state with n_{kσ}=0. A peculiarity of superconducting junctions is that internal relaxation processes cannot change the parity of the quasiparticle number. Parity changing processes require a quasiparticle from the bulk, and are therefore rare as the concentration of such quasiparticles is exponentially small at low temperatures. So one can say that a superconducting junction can be in two different ground states, with even or odd parity. The switching between parities occurs on a long time scale: experiments with break junctions^{21}, for example, yield switching times >0.1 ms, while for two other recent experimental setups 10 ms (ref. 22) and 1 min (ref. 23) have been reported. We therefore expect a switching of the Chern number and the associated quantized transconductance on this time scale. In particular, for the situation we concentrate on, the nontrivial Chern number C comes from the lowest band, and the transconductance would switch between −(4e^{2}/h)C for even parity and zero for odd parity. If the current is averaged over time intervals much longer than the switching scale, the resulting transconductance will be proportional to the probability of finding the junction in the even parity state.
Therefore, the topological signature is robust provided the fermion parity is preserved, a fact which has also been pointed out for other topological systems realized in superconductors^{9}. There are ways to control quasiparticle poisoning and reach a desired even parity in the ABS occupation^{7,24}. Note that quantum Halllike conductance quantization has also been proposed in superconducting devices with finite charging energy and hosting quantum phase slips^{25}. Furthermore, superconducting junctions with a gatetunable charging energy may realize topologically protected discrete charge pumping^{26}.
We now focus on a fourterminal junction and investigate the energy spectrum as a function of the three independent phases ϕ_{1,2,3}. As mentioned above, such a 3D bandstructure may host Weyl points with positive or negative topological charge. The NielsenNinomiya theorem^{27} implies that the total topological charge of the system is zero, such that the number of Weyl points is always even. Furthermore, timereversal invariance corresponds to a mapping from to , hence a Weyl point at has a counterpart at with the same topological charge. Thus, Weyl points come in groups of 4.
In the simplest case, where each contact contains only one channel, the system may realize 0 or 4 Weyl points, corresponding to a topologically trivial or nontrivial 3D material, respectively. If a scattering matrix yielding Weyl points is found, small changes in the scattering matrix only modify their position, but cannot gap them. Namely, as the Weyl points carry a topological charge, individual Weyl points are stable and annihilation is possible only when two Weyl points with opposite charges coincide.
A specific example is shown in Fig. 2. The position and charge of the 4 Weyl points is shown in Fig. 2a. Without loss of generality, we fix the phase ϕ_{3} and compute the transconductance G^{12} between voltagebiased contacts 1 and 2. In Fig. 2b, one can clearly see that the transconductance increases (decreases) by 4e^{2}/h when ϕ_{3} passes through a Weyl point with positive (negative) charge. Interpreting ϕ_{3} as a control parameter rather than a quasimomentum, we thus see that the 2D bandstructure of the system as a function of ϕ_{1} and ϕ_{2} undergoes a topological transition when ϕ_{3} passes through a Weyl point. The transconductance directly measures the Chern number characterizing the corresponding 2D topological phase. Note that the transconductance satisfies the relation G^{12}(−ϕ_{3})=−G^{12}(ϕ_{3}) due to timereversal symmetry.
By randomly generating scattering matrices from the circular orthogonal ensemble^{28}, we find that about 5% of scattering matrices give rise to four Weyl points (Supplementary Note 4). More Weyl points can be obtained in multichannel junctions where the maximal number of Weyl points is roughly proportional to the number of channels, and the probability to have no Weyl points is small (Supplementary Fig. 2). As a consequence, a greater variety of 2D topological phases with higher Chern numbers can be realized in that case. This is shown in Fig. 3 for a multichannel junction hosting 36 Weyl points, where the maximal Chern number is 3. Recently realized fewchannel cross junctions^{29} are promising to observe the transconductance in a fourterminal junction.
We now turn to fiveterminal junctions. In that case, the Weyl singularities appear as closed loops in the 4D space of phases. The simplest way to visualize them is to consider the additional phase ϕ_{4} as a tuning parameter of the 3D system described by the phases (ϕ_{1}, ϕ_{2}, ϕ_{3}). Tuning ϕ_{4} the Weyl points move, but remain at zero energy. Note that in the 3D subspace, timereversal symmetry is effectively broken, as for a fixed nonzero ϕ_{4}, a Weyl point at does not have a counterpart at anymore. The only constraint at a finite ϕ_{4} is that the number of Weyl points is even. Once two Weyl points with opposite charge meet, they annihilate. Thus their trajectories describe closed loops in the 4D space of all phases.
As before, the presence of Weyl singularities may be probed by the transconductance between two voltagebiased terminals, say terminal 1 and 2. As a function of the two other phases ϕ_{3} and ϕ_{4}, we now obtain areas with different quantized values of the transconductance, corresponding to different Chern numbers. The boundaries of these areas are given by the projections of the Weyl loops on the (ϕ_{3}, ϕ_{4})plane. An example is shown in Fig. 4. Here timereversal symmetry of the scattering matrix manifests itself in the relation G^{12}(−ϕ_{3}, −ϕ_{4})=−G^{12}(ϕ_{3}, ϕ_{4}).
Discussion
The above considerations are straightforwardly extended to a larger number of terminals. Furthermore, as we saw for the fiveterminal junction, breaking timereversal invariance does not lift the topological protection of the Weyl points at zero energy, as long as spin degeneracy is preserved. When spin degeneracy is lifted due to a Zeeman field or due to spin–orbit interactions, we expect the Weyl points to shift away from zero energy while remaining stable (Supplementary Note 5 and Supplementary Fig. 3). The possibility of realizing zeroenergy states in threeterminal junctions with strong spin–orbit interaction was studied in ref. 30.
To summarize, we predict the existence of topological Weyl singularities in the ABS spectrum of nterminal superconducting junctions with n≥4. These Weyl singularities manifest themselves in a quantized transconductance between two voltagebiased contacts, when the remaining phases are tuned away from the singularities. Tuning the system through a Weyl singularity, the conductance displays a step, signalling a topological transition of the 2D system described by the two phases of the voltagebiased contacts. This signature is robust provided the system remains in its ground state. Thus, the quantized transconductance should be accessible experimentally at low temperatures and voltages, in multiterminal junctions with, for example, 2D electron gases, semiconducing crossed nanowires or graphene (all systems through which a conventional Josephson effect has already been measured).
Additional information
How to cite this article: Riwar, R.P. et al. Multiterminal Josephson junctions as topological matter. Nat. Commun. 7:11167 doi: 10.1038/ncomms11167 (2016).
References
Josephson, B. D. Possible new effects in superconductive tunnelling. Phys. Lett. 1, 251–253 (1962).
Golubov, A. A., Kupriyanov, Yu. M., Kupriyanov & Il’ichev, E. The currentphase relation in Josephson junctions. Rev. Mod. Phys. 76, 411–469 (2004).
Devoret, M. H., Wallraff, A. & Martinis, J. M. Superconducting qubits: A short review, Preprint at http://arxiv.org/abs/0411174 (2004).
Beenakker, C. W. J. Universal limit of criticalcurrent fluctuations in mesoscopic Josephson junctions. Phys. Rev. Lett. 67, 3836–3839 (1991).
Della Rocca, M. L. et al. Measurement of the currentphase relation of superconducting atomic contacts. Phys. Rev. Lett. 99, 127005 (2007).
Pillet, J.D. et al. Andreev bound states in supercurrentcarrying carbon nanotubes revealed. Nat. Phys. 6, 965–969 (2010).
Bretheau, L., Girit, C. O., Pothier, H., Esteve, D. & Urbina, C. Exciting Andreev pairs in a superconducting atomic contact. Nature 499, 312–315 (2013).
Kitaev, A. Yu. Unpaired majorana fermions in quantum wires. Phys.Usp. 44, 131136 (2001).
Fu, L. & Kane, C. L. Josephson current and noise at a superconductor/quantumspinhallinsulator/superconductor junction. Phys. Rev. B 79, 161408 (2009).
Alicea, J. New directions in the pursuit of majorana fermions in solid state systems. Rep. Prog. Phys. 75, 076501 (2012).
Badiane, D. M., Glazman, L. I., Houzet, M. & Meyer, J. S. Ac Josephson effect in topological Josephson junctions. C R Phys. 14, 840856 (2013).
Herring, C. Accidental degeneracy in the energy bands of crystals. Phys. Rev. 52, 365–373 (1937).
Murakami, S. Phase transition between the quantum spin hall and insulator phases in 3D: emergence of a topological gapless phase. New J. Phys. 9, 356356 (2007).
Wan, X., Turner, A. M., Vishwanath, A. & Savrasov, S. Y. Topological semimetal and Fermiarc surface states in the electronic structure of pyrochlore iridates. Phys. Rev. B 83, 205101 (2011).
Burkov, A. A. & Balents, L. Weyl semimetal in a topological insulator multilayer. Phys. Rev. Lett. 107, 127205 (2011).
Hosur, P. & Qi, X. Recent developments in transport phenomena in Weyl semimetals. C R Phys. 14, 857–870 (2013).
Castro Neto, A. H., Guinea, F., Peres, N. M. R., Novoselov, K. S. & Geim, A. K. The electronic properties of graphene. Rev. Mod. Phys. 81, 109–162 (2009).
Matsuura, S., Chang, P.Y., Schnyder, A. P. & Ryu, S. Protected boundary states in gapless topological phases. New J. Phys. 15, 065001 (2013).
Zhao, Y. X. & Wang, Z. D. Topological classification and stability of Fermi surfaces. Phys. Rev. Lett. 110, 240404 (2013).
Zhang, F. & Kane, C. L. Anomalous topological pumps and fractional Josephson effects. Phys. Rev. B 90, 020501 (2014).
Zgirski, M. et al. Evidence for longlived quasiparticles trapped in superconducting point contacts. Phys. Rev. Lett. 106, 257003 (2011).
Higginbotham, A. P. et al. Parity lifetime of bound states in a proximitized semiconductor nanowire. Nat. Phys. 11, 1017–1021 (2015).
van Woerkom, D. J., Geresdi, A. & Kouwenhoven, L. P. One minute parity lifetime of a NbTiN Cooperpair transistor. Nat. Phys. 11, 547–550 (2015).
Riwar, R.P., Houzet, M., Meyer, J. S. & Nazarov, Yu. V. Control of Andreev bound state population and related chargeimbalance effect. J. Phys. Condens. Matter 27, 095701 (2015).
Hriscu, A. M. & Nazarov, Yu. V. Quantum synchronization of conjugated variables in a superconducting device leads to the fundamental resistance quantization. Phys. Rev. Lett. 110, 097002 (2013).
Leone, R., Levy, L. P. & Lafarge, P. Cooperpair pump as a quantized current source. Phys. Rev. Lett. 100, 117001 (2008).
Nielsen, H. B. & Ninomiya, M. Absence of neutrinos on a lattice: (i). Proof by homotopy theory. Nucl. Phys. B 185, 20–40 (1981).
Mehta, M. L. Random Matrices Vol. 142, Elsevier/Academic Press (2004).
Plissard, S. R. et al. Formation and electronic properties of InSb nanocrosses. Nat. Nanotechnol. 8, 859–864 (2014).
van Heck, B., Mi, S. & Akhmerov, A. R. Single fermion manipulation via superconducting phase differences in multiterminal Josephson junctions. Phys. Rev. B 90, 155450 (2014).
Acknowledgements
This work has been supported by the Nanosciences Foundation in Grenoble, in the frame of its Chair of Excellence program. M.H. and J.M. acknowledge support through grants No. ANR11JS0400301 and ANR12BS04001603, and an EUFP7 Marie Curie IRG.
Author information
Authors and Affiliations
Contributions
M.H., J.S.M., and Y.V.N. formulated the model. R.P.R. and Y.V.N. performed the analytical derivations. R.P.R. conducted the numerical calculations. All authors contributed equally to design and writing of the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing financial interests.
Supplementary information
Supplementary Information
Supplementary Figures 13, Supplementary Notes 15 and Supplementary References (PDF 465 kb)
Rights and permissions
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
About this article
Cite this article
Riwar, RP., Houzet, M., Meyer, J. et al. Multiterminal Josephson junctions as topological matter. Nat Commun 7, 11167 (2016). https://doi.org/10.1038/ncomms11167
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/ncomms11167
 Springer Nature Limited
This article is cited by

Phaseengineering the Andreev band structure of a threeterminal Josephson junction
Nature Communications (2023)

Quasiperiodic circuit quantum electrodynamics
npj Quantum Information (2023)

Intermediate states in Andreev bound state fusion
Communications Physics (2023)

Gatetunable superconducting diode effect in a threeterminal Josephson device
Nature Communications (2023)

Compact description of quantum phase slip junctions
npj Quantum Information (2023)