Abstract
Topological physics relies on Hamiltonian’s eigenstate singularities carrying topological charges, such as Dirac points, and – in non-Hermitian systems – exceptional points (EPs), lines or surfaces. So far, the reported non-Hermitian topological transitions were related to the creation of a pair of EPs connected by a Fermi arc out of a single Dirac point by increasing non-Hermiticity. Such EPs can annihilate by reducing non-Hermiticity. Here, we demonstrate experimentally that an increase of non-Hermiticity can lead to the annihilation of EPs issued from different Dirac points (valleys). The studied platform is a liquid crystal microcavity with voltage-controlled birefringence and TE-TM photonic spin-orbit-coupling. Non-Hermiticity is provided by polarization-dependent losses. By increasing the non-Hermiticity degree, we control the position of the EPs. After the intervalley annihilation, the system becomes free of any band singularity. Our results open the field of non-Hermitian valley-physics and illustrate connections between Hermitian topology and non-Hermitian phase transitions.
Similar content being viewed by others
Introduction
So far, topological physics has been mostly dealing with Hermitian Hamiltonians, possessing well-defined topological invariants, such as the Chern number, calculated from the eigenstates of these Hamiltonians1. The topological charges composing these invariants are associated with the Hamiltonian singularities in the parameter space2, such as Dirac points. The development of non-Hermitian physics brought about new topological invariants3, whose relation to the Hermitian ones is a particularly active topic4.
The eigenstates of a non-Hermitian Hamiltonian are, in general, non-orthogonal. Exceptional points (EPs), where the eigenstates coalesce, can appear at the maxima of non-orthogonality when the non-Hermiticity is increased. EPs are known in optics for more than a century5, but only recently they have been shown to allow remarkable phenomena6, such as specific lasing7, unidirectional transport8, enhanced sensing9,10, or scattering control11. Their importance has been revealed thanks to their description in terms of a topological charge, characterizing a topological phase3,12,13,14, which can be measured by encircling the EP in either parameter15 or reciprocal space4. EPs always appear in pairs connected by a Fermi arc in the full parameter space, similar to the Weyl points in Hermitian systems, as described by the famous Nielsen-Ninomiya no-go theorem16. Each pair of EPs is formed from a minimum of the Hermitian coupling (e.g. band crossing).
So far, the reported non-Hermitian topological transitions were related to the creation of EPs with the increase of non-Hermiticity4,17,18,19,20. Typically, one Dirac point (DP) splits into two EPs by increasing non-Hermiticity (Fig. 1a). These EPs can then annihilate by reducing non-Hermiticity (going back to the Hermitian limit). This transition converts a Hermitian singularity into two EPs and vice versa. More complicated situations can occur when more than two levels or bands get coupled, which leads to the emergence of higher-order singularities21,22. These non-Hermitian transitions are based on local (geometrical) properties of states in parameter space. They do not depend on the possible presence of other DPs, and on the global geometry (topology) of eigenstates.
In this work, we demonstrate a different type of non-Hermitian topological transition in a continuous (non-periodic) 2D photonic system. We show that, if the Hermitian Hamiltonian is topologically trivial, supporting opposite-sign DPs, an EP issued from a DP can be moved towards another EP issued from another DP with which it annihilates (Fig. 1b). This process takes place upon the increase of non-Hermiticity and there is no singularity of any type (neither EPs nor DPs) left after the annihilation. It relies not only on the existence of an isolated DP, but on the global band topology, which takes into account not just one, but all singularities present in the parameter space.
Results
Experimental system and its model
The actual platform we study is composed of two microcavities filled with liquid crystal23,24 (Fig. 2a). These microcavities (see Methods for structural details) host a series of photonic modes with quantized wavevectors perpendicular to the mirror plane and energies EN (N is the mode number). Each mode forms a polarization doublet showing an in-plane parabolic dispersion with a 2D effective mass mN ∼ N. The polarization degeneracy in a doublet is lifted at all wavevectors except k = 0 (touching parabolas) by the splitting between TE and TM eigenmodes (Transverse-Electric and Transverse-Magnetic). This splitting acts as photonic spin-orbit coupling (SOC) characterized by a winding number 225,26. The liquid crystal molecules orientation is set by an external voltage, which controls the linear birefringence α23,24. A small α < (EN+1 − EN) lifts the k = 0 degeneracy. The crossing between the two modes of same order ((N,N)-case) leads to the formation of two tilted Dirac cones both carrying the same topological Berry charge + 1/226. When α becomes comparable with (EN+1 − EN), modes of different parities become energetically close and get coupled by a Rashba-Dresselhaus SOC with equal strength23, also called emergent optical activity27. Here we consider α ≈ (EN+2 − EN)24, so that this optical activity is negligible. The corresponding eigenmodes look like two 2D parabola (Fig. 2b). Neglecting the losses, these two bands can be described by the following effective 2 × 2 Hermitian Hamiltonian written on the circular polarization basis:
where \({E}_{{{{{{{{\rm{H}}}}}}}}}^{N+2}\) and mH ∼ N + 2 are the energy and mass of the N + 2th H-polarized mode, and \({E}_{{{{{{{{\rm{V}}}}}}}}}^{N}\) and mV ∼ N are the energy and mass of the V-polarized mode number N. kx, ky are the 2D wavevector components. The spin-independent masses mx and my are determined by the birefringence and the angle of the optical axis (see Supplementary Note I). β is the magnitude of the TE-TM SOC, \(\beta ^{\prime}={\hslash }^{2}({m}_{{{{{{{{\rm{H}}}}}}}}}-{m}_{{{{{{{{\rm{V}}}}}}}}})/4{m}_{{{{{{{{\rm{H}}}}}}}}}{m}_{{{{{{{{\rm{V}}}}}}}}}\) and \({{\Delta }}=({E}_{{{{{{{{\rm{H}}}}}}}}}^{N+2}-{E}_{{{{{{{{\rm{V}}}}}}}}}^{N})/2\). This Hermitian Hamiltonian can be written as a linear combination of identity and Pauli matrices, which defines a real effective magnetic field Ωr acting on the polarization pseudospin (H = Ωr ⋅ S). The two non-zero components of the field are \({{{\Omega }}}_{{{{{{{{\rm{r}}}}}}}}}^{x}={{\Delta }}-\beta ^{\prime} {k}^{2}-\beta ({k}_{x}^{2}-{k}_{y}^{2})\) and \({{{\Omega }}}_{{{{{{{{\rm{r}}}}}}}}}^{y}=-2\beta {k}_{x}{k}_{y}\).
Hermitian topology
This effective Hamiltonian possesses two distinct topological phases we experimentally characterize below (see also Supplementary Notes I and II for more details). If \(\beta \, > \, \beta ^{\prime}\), the bands show two tilted DPs carrying the same topological charge (Fig. 2i), and the bands are topologically non-trivial, characterized by a non-zero winding of the pseudospin giving rise to a non-zero Chern number, if a gap is opened by breaking the time-reversal symmetry. If \(\beta \, < \beta^{\prime}\), the bands possess four tilted Dirac cones, as shown in Fig. 2c, f. Their coordinates are given by ( ± k0x, 0) and (0, ± k0y), where \({k}_{0,x,y}={({{\Delta }}/(\beta ^{\prime} \pm \beta ))}^{1/2}\). The winding number of the pseudospin is + 1 for each of the two DPs located on the ky-axis, which corresponds to Berry curvature monopoles of charge + 1/2. The DPs located on the kx-axis carry a pseudospin winding number − 1 and a Berry curvature charge − 1/2 (Fig. 2f). The corresponding bands are therefore globally topologically trivial, with zero overall pseudospin winding and a vanishing Chern number.
Figure 3 demonstrates the above-mentioned Hermitian topological transition. It presents polarisation pseudospin (see Supplementary Note I) textures, experimentally extracted from a microcavity region with a liquid crystal layer thickness of around 3.9 μm through polarization-resolved tomography. Figure 3a is observed with 1.39 V applied to ITO (Indium Tin Oxide) electrodes. At this voltage, the modes N + 2 and N cross each other along both wave vector directions. Four pseudospin monopoles corresponding to Dirac points are observed. The total winding of the pseudospin encircling all four Dirac points is zero, as can be seen from the high-k texture (pseudospin pointing to the right). Under a different external voltage of 11 V, modes with the numbers N and N are almost degenerate. The pseudospin texture shown in Fig. 3b exhibits two monopoles, whose position along either the kx or ky axis is controlled by the sign of Δ. In that case, the total winding number of the pseudospin is 2 (double winding stemming from the TE-TM field).
Non-Hermitian transition
The next step is to consider the losses (line broadening), inherently present in photonic systems. Importantly, in the (N+2, N) case, these losses are significantly different for the H and V modes with ΓH = 2.04 ± 0.04 meV, ΓV = 1.8 ± 0.1 meV (see Methods and Supplementary Note IV). This requires adding a non-Hermitian part to the total Hamiltonian of the system, as already illustrated in4,18,19:
where \({H}^{imag}=i({{{\Gamma }}}_{0}{{\mathbb{I}}}_{2}+\delta {{\Gamma }}{\sigma }_{x})\) with Γ0 = (ΓH + ΓV)/2 being the mean decay rate. The term δΓ = (ΓH − ΓV)/2 defines a constant imaginary effective field along x: Ωi = (δΓ, 0, 0)T. As shown in Fig. 2d, g, this non-Hermitian part transforms each DP into a pair of EPs5,19,28 connected by a line, called a Fermi arc14, where the real parts of the eigenvalues are degenerate. The squared absolute value of the complex splitting between the eigenmodes reads \(4(|{{{{{{{{\mathbf{\Omega }}}}}}}}}_{{{{{{{{\rm{r}}}}}}}}}^{2}-{{{{{{{{\mathbf{\Omega }}}}}}}}}_{{{{{{{{\rm{i}}}}}}}}}^{2}{|}^{2}+4|{{{{{{{{\mathbf{\Omega }}}}}}}}}_{{{{{{{{\rm{r}}}}}}}}}{{{{{{{{\mathbf{\Omega }}}}}}}}}_{{{{{{{{\rm{i}}}}}}}}}{|}^{2})\). The existence of an EP (zero splitting) therefore requires ΩrΩi = 0 and \({{{{{{{{\mathbf{\Omega }}}}}}}}}_{{{{{{{{\rm{r}}}}}}}}}^{2}-{{{{{{{{\mathbf{\Omega }}}}}}}}}_{{{{{{{{\rm{i}}}}}}}}}^{2}=0\). In our case, the first condition reads:
which determines an ellipse of possible locations for EPs (Fig. 2g, cyan and magenta points). The second condition \({{{{{{{{\mathbf{\Omega }}}}}}}}}_{{{{{{{{\rm{r}}}}}}}}}^{2}-{{{{{{{{\mathbf{\Omega }}}}}}}}}_{{{{{{{{\rm{i}}}}}}}}}^{2}=0\) is verified along the blue curves in Fig. 2g, h. The crossing of both lines sets the coordinates of the 8 EPs:
The Fermi arc, shown in blue-green, is a part of the ellipse where \({{{{{{{{\mathbf{\Omega }}}}}}}}}_{{{{{{{{\rm{r}}}}}}}}}^{2}-{{{{{{{{\mathbf{\Omega }}}}}}}}}_{{{{{{{{\rm{i}}}}}}}}}^{2} \, < \, 0\), whereas the other part, given by \({{{{{{{{\mathbf{\Omega }}}}}}}}}_{{{{{{{{\rm{r}}}}}}}}}^{2}-{{{{{{{{\mathbf{\Omega }}}}}}}}}_{{{{{{{{\rm{i}}}}}}}}}^{2} \, > \, 0\) and shown in gray is the imaginary Fermi arc14 with degenerate imaginary parts of the modes.
The topological charge of an EP can be defined as the winding number of the complex energy of eigenmodes around the EP3,4,12,13:
The winding numbers of the eight EPs alternate in sign along the ellipse. Increasing δΓ (or decreasing Δ) increases the degree of non-Hermiticity and moves the EPs away from the spawning points, along the ellipse, until they meet each other and annihilate, as shown in Fig. 2e, h. Fermi arcs connect to form a closed line of trivial degeneracy. It is not a ring of exceptional points reported in29, because the imaginary parts of the energies are not degenerate along this whole line.
As shown above (Fig. 3), a topological transition occurs for the Hermitian part of the Hamiltonian when \(\beta ^{\prime}=\beta\). The system switches between four DPs (globally trivial) and two DPs (non-trivial). In the case with 2 DPs, k0y becomes imaginary. Equation (3) determining the location of EPs remains valid, but describes two hyperbolas (Fig. 2j, k). EPs issued from distinct DPs are moving towards infinity on separate open curves and cannot meet anymore. The EP annihilation cannot occur (See Supplementary Note VI for a detailed discussion).
Going back to the case \(\beta \, < \, \beta ^{\prime}\), the EP annihilation occurs in Eqs. (5) when δΓ = βk0xk0y, which gives
The degree of non-Hermiticity χ can be changed either by increasing δΓ or decreasing Δ. The latter option is used in our experiment: Δ is controlled by the voltage affecting the liquid crystal molecules orientation. The experimental study of the liquid crystal cavity is performed by polarization-resolved transmission, from which we extract the real and imaginary parts of the energies of the eigenmodes and also their polarization (see Methods). The differences between the real (imaginary) part of the energies versus kx, ky are shown in Fig. 4a, e and c,g for two values of detuning 2Δ = 3 meV and 1.2 meV for the cases where the EPs are present and annihilated, respectively. In Fig. 4a, e, one can observe the Fermi arcs, where the real parts of energy are degenerate and the imaginary parts split. The EPs are shown by white crosses. On the other hand, imaginary Fermi arcs are the lines where the imaginary parts of energies are degenerate and the real parts split. The real and imaginary part of energy, and their uncertainties (see Methods and Supplementary Note V) along the Fermi arc with an EP are shown in Fig. S6. After the EPs annihilation, only a real Fermi arc remains (Fig. 4c, g). Figure 4b, d, f, h shows the corresponding theoretical results, obtained using the effective non-Hermitian Hamiltonian (2). The agreement between experiment and theory is excellent. We note that in ref. 24, the very same structure was studied in the regime of crossing between the N + 2 and N modes. The key difference is that the experiment was performed at higher detuning Δ, so with a very small non-Hermiticity degree. This is also the regime for which Fig. 3 was measured. In that case Fermi arcs are very small, the EPs could not be resolved and an Hermitian description of bands is appropriate.
The topological charge measurement is presented in Fig. 4i–l. Fig. 4i, k show a map of the phase of the complex energy of the lower mode for 2Δ = 3 meV and 2Δ = 1.2 meV, as previously. In both cases, the real Fermi arcs appear as a sharp phase shift. Very clear phase vortices are visible at the EP positions in Fig. 4i and are absent in Fig. 4k. These features are in excellent agreement with the simulations based on the effective non-Hermitian Hamiltonian (2), shown in Fig. 4j, l. The topological charge of each EP w = ± 1/2 is determined by the direction of the phase vortex winding. These charges are opposite for the EPs originating from different DPs, which ultimately allows their annihilation.
Figure 4m demonstrates the control of the Fermi arc angular size related to the EPs position through the experimental tuning of the non-Hermiticity χ. The topological transition associated with the EP annihilation is clearly visible taking place for χ = 1. To demonstrate that the observed behavior does not depend on a particular sample, we have performed extra measurements with a sample characterized by different parameters. Black and red dots in panel 4m correspond to samples 1 and 2 respectively, whereas all data shown in panels 4(a–l) are from sample 1. The theoretical curve is universal, it does not have any fitting parameters.
Discussion
Non-Hermitian transitions in two-band systems through EP merging are typically related to a single Hermitian singularity3,13,14,18. Here, we consider the merging of EPs originating from different DPs upon increasing the non-Hermiticity, which to our knowledge was not reported before. It can be viewed as a first example of non-Hermitian multivalley physics and demonstrates the link between Hermitian topology and non-Hermitian phase transitions. Indeed, there is no singularity of any type after the transition, as in the transition with the annihilation of 2D DPs carrying opposite charges30. The phase transition we observe could be realized in other multivalley systems, like artificial graphene with a σx non-Hermitian contribution.
From an applied perspective, our work sets microcavities alongside the waveguide-based photonic systems31 as a reconfigurable platform for exploring non-Hermitian topology. We demonstrate the tuning of the EP coordinates in k-space by simple modification of an external voltage, in a micro-device, at optical frequencies. This could allow to control the angle of emission of the modes surrounding the EPs, which are known to possess remarkable properties6. Another interest of the planar cavity platform is that it allows implementing interacting photons modes (exciton-polaritons)4 possibly up to the recently demonstrated single-photon non-linearity32. These possibilities combined with our present finding could allow to address in future non-Hermitian topological physics for strongly interacting-correlated particles.
Methods
Samples
Both cavities consist of two distributed Bragg reflectors made of 6 SiO2/TiO2 pairs with maximum reflectance at 550 nm grown on glass plates with ITO electrodes. Space between the DBRs is filled with highly birefringent liquid crystal with Δn = 0.41. To obtain homogeneous orientation of LC, both DBRs are finished with structured polymer orienting layer. The cavity 1 studied in Figs. 4, (S4–S7) has a total thickness of the LC layer of approx. 1.8 μm. The cavity 2 studied in Figs. 3, 4m and Fig. S8 has a thickness of 3.2 μm.
Experimental setup
Experimental results were obtained in transmission configuration. Broadband light from a LED diode was focused on the sample with a microscope objective with 50 × magnification and numerical aperture NA = 0.6. Transmitted light was collected by another objective with 20 × magnification and NA = 0.4. Fourier plane of the collecting objective was imaged on the entrance slit of a monochromator equipped with a CCD camera. Both wave vector directions were measured by scanning of the image across the slit by the automated movement of the imaging lens. Data was collected independently for 6 incident light polarizations; linear: horizontal, vertical, diagonal, antidiagonal and circular σ+, σ− by adjusting angles of half wave plate and quarter wave plate after fixed linear polarizer. LC layer anisotropy in x–z plane was controlled by external square waveform applied to ITO electrodes with 1 kHz frequency and amplitude of 1.77 V (Fig. 4a, e, i) and 1.72 V (Fig. 4c, g, k).
Linewidth extraction
We extract the real and imaginary parts of the energies (that is, the positions and the linewidths) of the modes from the polarization-resolved spectra by fitting them with the Voigt function, in order to account both for the homogeneous broadening due to the mode lifetime and for the inhomogeneous broadening due to disorder. Only the homogeneous part of the broadening (Lorentzian linewidth) can give rise to non-Hermiticity and is accounted for by the Hamiltonian (2). Parallel computing is used to speed up the extraction for the whole reciprocal space with high resolution. An example of the energy spectrum in two polarizations, together with its fit, is provided in the Supplementary Note IV.
The variation of the difference of the linewidths δΓ with detuning Δ is negligible, because the linewidths scale as the energies of the modes and their difference scales as the birefringence Δn, responsible for the difference of the energies between the modes of the same order \({E}_{{{{{{{{\rm{H}}}}}}}}}^{N}-{E}_{{{{{{{{\rm{V}}}}}}}}}^{N}\), whereas the detuning Δ between the almost degenerate modes \({E}_{{{{{{{{\rm{H}}}}}}}}}^{N+2}-{E}_{{{{{{{{\rm{V}}}}}}}}}^{N}\) does not account for the whole birefringence, but only for a small part of it. Indeed, in our case the overall splitting due to the birefringence \({E}_{{{{{{{{\rm{H}}}}}}}}}^{N}-{E}_{{{{{{{{\rm{V}}}}}}}}}^{N} \sim 200\) meV, and its change with voltage (corresponding to the variation of Δ) is of the order of 1 meV. The variation of δΓ with voltage is of the same order as that of the birefringence (about 1%), which is smaller than the experimental uncertainty of this parameter (20%).
The uncertainty on the linewidth is the main source of uncertainty on the non-Hermiticity degree χ, used to plot Fig. 4m. Other non-negligible contributions come from the uncertainty on the polarization splittings β and \(\beta ^{\prime}\). As to the length of the Fermi arcs, we have used 3 arcs out of 4 for averaging for sample 1, since the 4th one (the one at ky < 0) is systematically affected by experimental measurement problems. The lengths of the remaining 3 arcs are very close (e.g. 36, 37, and 38 degrees for the experiment presented in Fig. 4a). For sample 2, we have used all 4 arcs, and the uncertainty is slightly larger. The uncertainty on the position of the EPs is relatively low, because this position is constrained by 2 independent measurements and a third one combining them (real and imaginary parts of the energy, and its phase).
Experimental uncertainty and the evidence for topological transition
The 2D images shown in Fig. 4 of the main text do not allow to indicate the uncertainties, which are important to prove that the transition associated with the annihilation of the EPs indeed takes place. To show this, we plot in Fig. S6 the real and imaginary part of energy, and their uncertainties along the Fermi arc when an EP is present. We also show in Fig S7 a cross-section of Fig. 4a, d of the main text near the top right corner of the Fermi arc denoted by \(k^{\prime}\) to clearly show the difference between the case where EPs exist and when they annihilate. In one case, this cross-section crosses an imaginary Fermi arc, and in the other case – a real Fermi arc that forms a full circle.
Parameters of the Hamiltonian
The theoretical panels of Fig. 4 were calculated using the following parameters, obtained from the sample 1 by fitting the dispersion (Supplementary Note III) and from the linewidth extraction discussed above: mx = (1.34 ± 0.07) × 10−5m0, my = (1.08 ± 0.06) × 10−5m0, β = 0.080 ± 0.03 meVμm2, \(\beta ^{\prime}=0.47\pm 0.03\) meVμm2. Other parameters were already given in the text, but we provide them here for convenience: 2Δ = 2.7 ± 0.1 and 1.2 ± 0.1 meV, 2δΓ = 0.24 ± 0.05 meV.
The parameters of the sample 2 are given here for comparison: mx = (1.27 ± 0.03) × 10−5m0, my = (1.14 ± 0.03) × 10−5m0, β = 0.23 ± 0.01 meVμm2, \(\beta ^{\prime}=0.35\pm 0.01\) meVμm2, 2δΓ = 0.5 ± 0.07 meV.
Data availability
The data generated in this study are available in the Open Science Framework (OSF) repository: https://osf.io/jnx8k/?view_only=16426f9a35404264badaaa93162060a7.
References
Hasan, M. Z. & Kane, C. L. Colloquium: Topological insulators. Rev. Mod. Phys. 82, 3045–3067 (2010).
Mera, B. & Ozawa, T. Kähler geometry and chern insulators: relations between topology and the quantum metric. Phys. Rev. B 104, 045104 (2021).
Leykam, D., Bliokh, K. Y., Huang, C., Chong, Y. D. & Nori, F. Edge modes, degeneracies, and topological numbers in non-Hermitian systems. Phys. Rev. Lett. 118, 040401 (2017).
Su, R. et al. Direct measurement of a non-Hermitian topological invariant in a hybrid light-matter system. Sci. Adv. 7, eabj8905 (2021).
Voigt, W. On the behaviour of pleochroitic crystals along directions in the neighbourhood of an optic axis. Philos. Mag. Series 4, 90 (1902).
Miri, M. A. & Alù, A. Exceptional points in optics and photonics. Science 363 (2019).
Peng, B. et al. Loss-induced suppression and revival of lasing. Science 346, 328 (2014).
Doppler, J. et al. Dynamically encircling an exceptional point for asymmetric mode switching. Nature 537, 76–79 (2016).
Hodaei, H. et al. Enhanced sensitivity at higher-order exceptional points. Nature 548, 187–191 (2017).
Chen, W., Özdemir, Ş. K., Zhao, G., Wiersig, J. & Yang, L. Exceptional points enhance sensing in an optical microcavity. Nature 548, 192–196 (2017).
Song, Q., Odeh, M., Zúñiga-Pérez, J., Kanté, B. & Genevet, P. Plasmonic topological metasurface by encircling an exceptional point. Science 373, 1133–1137 (2021).
Gong, Z. et al. Topological phases of non-Hermitian systems. Phys. Rev. X 8, 031079 (2018).
Shen, H., Zhen, B. & Fu, L. Topological band theory for non-Hermitian Hamiltonians. Phys. Rev. Lett. 120, 146402 (2018).
Bergholtz, E. J., Budich, J. C. & Kunst, F. K. Exceptional topology of non-Hermitian systems. Rev. Modern Phys. 93, 015005 (2021).
Gao, T. et al. Observation of non-Hermitian degeneracies in a chaotic exciton-polariton billiard. Nature 526, 554–558 (2015).
Nielsen, H. & Ninomiya, M. A no-go theorem for regularizing chiral fermions. Phys. Lett. B 105, 219–223 (1981).
Zeuner, J. M. et al. Observation of a topological transition in the bulk of a non-Hermitian system. Phys. Rev. Lett. 115, 040402 (2015).
Zhou, H. et al. Observation of bulk fermi arc and polarization half charge from paired exceptional points. Science 359, 1009–1012 (2018).
Richter, S. et al. Voigt exceptional points in an anisotropic ZnO-based planar microcavity: Square-root topology, polarization vortices, and circularity. Phys. Rev. Lett. 123, 227401 (2019).
Ozturk, F. et al. Observation of a non-Hermitian phase transition in an optical quantum gas. Science 372, 88 (2021).
Ding, K., Ma, G., Xiao, M., Zhang, Z. & Chan, C. T. Emergence, coalescence, and topological properties of multiple exceptional points and their experimental realization. Phys. Rev. X 6, 021007 (2016).
Weiyuan, T. et al. Exceptional nexus with a hybrid topological invariant. Science 363, eaar7709 (2020).
Rechcińska, K. et al. Engineering spin-orbit synthetic Hamiltonians in liquid-crystal optical cavities. Science 366, 727–730 (2019).
Król, M. et al. Observation of second-order meron polarization textures in optical microcavities. Optica 8, 255–261 (2021).
Kavokin, A., Malpuech, G. & Glazov, M. Optical spin Hall effect. Phys. Rev. Lett. 95, 136601 (2005).
Gianfrate, A. et al. Measurement of the quantum geometric tensor and of the anomalous Hall drift. Nature 578, 381–385 (2020).
Jiahuan, R. et al. Nontrivial band geometry in an optically active system. Nat. Commun. 12, 1 (2021).
Liao, Q. et al. Experimental measurement of the divergent quantum metric of an exceptional point. Phys. Rev. Lett. 127, 107402 (2021).
Zhen, B. et al. Spawning rings of exceptional points out of Dirac cones. Nature 525, 354–358 (2015).
Tarruell, L., Greif, D., Uehlinger, G., Jotzu, G. & Esslinger, T. Creating, moving and merging Dirac points with a Fermi gas in a tunable honeycomb lattice. Nature 483, 302 (2012).
Wang, K. et al. Generating arbitrary topological windings of a non-hermitian band. Science 371, 1240–1245 (2021).
Zasedatelev, A. V. et al. Single-photon nonlinearity at room temperature. Nature 597, 493 (2021).
Acknowledgements
We acknowledge useful discussions with C. Leblanc. This work was supported by the National Science Centre grants 2019/35/B/ST3/04147, 2019/33/B/ST5/02658, 2018/31/N/ST3/03046 and 2017/27/B/ST3/00271, and the Ministry of National Defense Republic of Poland Program – Research Grant MUT Project 13-995 and MUT University grant (UGB) for the Laboratory of Crystals Physics and Technology for year 2021, and the European Union’s Horizon 2020 program, through a FET Open research and innovation action under the grant agreements No. 899141 (PoLLoC) and No. 964770 (TopoLight). We also acknowledge the support of the ANR Labex Ganex (ANR-11-LABX-0014), and of the ANR program "Investissements d’Avenir" through the IDEX-ISITE initiative 16-IDEX-0001 (CAP 20-25). P.G.L. acknowledges the support of the UK’s Engineering and Physical Sciences Research Council (grant EP/M025330/1 on Hybrid Polaritonics), the support of the RFBR project No. 20-52-12026 (jointly with DFG) and No. 20-02-00919.
Author information
Authors and Affiliations
Contributions
J.S., G.M., D.D.S., W.P., B.P., and P.G.L. acquired funding; M.Kr., P.O., M.Kę., M.M. and K.Ł-M. performed the experiments under the guidance of B.P. and J.S. with strong support from P.G.L.; P.K. synthesized liquid crystal; R.Ma., P.M., and W.P. constructed and fabricated the LC microcavity; I.S. performed theoretical analysis under the guidance of G.M. and D.D.S.; G.M. and D.D.S. proposed the Hamiltonian of the system with the non-Hermitian part; M.Kr. and W.B. developed numerical Berreman method modeling. All authors participated in the interpretation of experimental data; J.S., D.D.S., G.M., and W.P. supervised the project; I.S., M.Kr., G.M., D.D.S. wrote the manuscript with input from all other authors. M.Kr., B.P., and I.S. proposed the visualization of experimental and theoretical data.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks the anonymous, reviewer(s) for their contribution to the peer review of this work. Peer review reports are available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Król, M., Septembre, I., Oliwa, P. et al. Annihilation of exceptional points from different Dirac valleys in a 2D photonic system. Nat Commun 13, 5340 (2022). https://doi.org/10.1038/s41467-022-33001-9
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-022-33001-9
This article is cited by
-
Reconfigurable quantum fluid molecules of bound states in the continuum
Nature Physics (2024)
-
Non-Hermitian chiral degeneracy of gated graphene metasurfaces
Light: Science & Applications (2023)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.