The quantum behaviour of electrons in materials is the foundation of modern electronics and information technology1,2,3,4,5,6,7,8,9,10,11, and quantum materials with topological electronic and optical properties are essential for realizing quantized electronic responses that can be used for next generation technology. Here we report the first observation of topological quantum properties of chiral crystals6,7 in the RhSi family. We find that this material class hosts a quantum phase of matter that exhibits nearly ideal topological surface properties originating from the crystals’ structural chirality. Electrons on the surface of these crystals show a highly unusual helicoid fermionic structure that spirals around two high-symmetry momenta, indicating electronic topological chirality. The existence of bulk multiply degenerate band fermions is guaranteed by the crystal symmetries; however, to determine the topological invariant or charge in these chiral crystals, it is essential to identify and study the helicoid topology of the arc states. The helicoid arcs that we observe on the surface characterize the topological charges of ±2, which arise from bulk higher-spin chiral fermions. These topological conductors exhibit giant Fermi arcs of maximum length (π), which are orders of magnitude larger than those found in known chiral Weyl fermion semimetals5,8,9,10,11. Our results demonstrate an electronic topological state of matter on structurally chiral crystals featuring helicoid-arc quantum states. Such exotic multifold chiral fermion semimetal states could be used to detect a quantized photogalvanic optical response, the chiral magnetic effect and other optoelectronic phenomena predicted for this class of materials6.


The discovery of topological insulators has inspired the search for a wide variety of topological conductors1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26. One example of a topological conductor is the Weyl semimetal, which features emergent Weyl fermions as low-energy excitations of electrons. These Weyl fermions are associated with topological chiral charges (Chern numbers) that are located at twofold-degenerate band crossings1,4,5,8,9,11,12,13,14 in momentum space. Besides twofold-degenerate Weyl fermions4,5,8,9,10,11,12,13,14,15, various types of emergent topological chiral fermion may appear, including Kramers–Weyl fermions6,7 and higher-fold band fermions16,17,18. Recently, a few non-centrosymmetric crystals were identified in which a band inversion gives rise to a Weyl semimetal state4,8,9,10,11,20,21,22,23. However, all of these materials suffer from several drawbacks, such as large numbers of Weyl fermions, Weyl fermions close to each other in momentum space, and short Fermi arcs, which are much (orders of magnitude) less topologically robust than the ideal case allowed by theory. To thoroughly explore and utilize the robust and unusual quantum phenomena induced by chiral fermions in optics or magnetotransport, topological conductors with near-ideal electronic properties or new types of topological conductor are needed1,7.

A different approach towards searching for ideal topological conductors is to examine crystal symmetries, which can also lead to topological band crossings2,3,8. For instance, it has been shown that non-symmorphic symmetries can guarantee the existence of band crossings for certain electron fillings16,19,24. As another example we might consider structurally chiral crystals, the lattices of which possess no inversion, mirror or roto-inversion symmetries6. Structurally chiral crystals are expected to host a variety of topological band crossings, which are guaranteed to be pinned to time-reversal-invariant momenta6,25. Moreover, such crystals naturally give rise to a quantized circular photogalvanic current, the chiral magnetic effect and other novel transport and optics effects that are forbidden in known topological conductors, such as TaAs8,9,20,21,22,23.

Incorporating these paradigms into a broader search, we have studied various candidate nonmagnetic and magnetic conductors, such as MoTe2, WTe2, Mn3Ge, Ge3Sn, Mn3Sn, Na3Bi, GdPtBi, the LuPtSb family, HgCr2Se4, pyrochlore iridates, LaPtBi, Co3Sn2S2, Fe3Sn2 and the XSi (X = Rh, Co) family, with advanced spectroscopic techniques. Many of these materials exhibit either large co-existing trivial bulk Fermi surfaces or surface reconstruction masking topological states. As often is the case with surface-sensitive techniques, the experimentally realized surface potential associated with a cleaved crystal may or may not allow these unusual electronic states to be observed. Thus, despite the new search paradigms, the discovery of topological materials suitable for spectroscopic experiments has remained a considerable challenge. Of all the materials that we explored, we observed that the XSi family of chiral crystals comes close to the experimental realization of the sought-after ideal topological conductor. Additionally, it is a new type of topological conductor beyond Weyl semimetals. Here we report high-resolution angle-resolved photoemission spectroscopy (ARPES) measurements in combination with state-of-the-art ab initio calculations to demonstrate new topological chiral properties in CoSi and RhSi. These chiral crystals approach ideal topological conductors because of their large Fermi arcs, which is related to the fact that they host the minimum non-zero number of chiral fermions—topological properties that we observe experimentally for the first time, to our knowledge.

The XSi family of materials crystallizes in a structurally chiral cubic lattice with space group P213, number 198 (Fig. 1a, Extended Data Figs. 1, 2). We confirmed the chiral crystal structure of our CoSi samples using single-crystal X-ray diffraction (Fig. 1b, Extended Data Fig. 3, Extended Data Table 1) and the associated three-dimensional Fourier map (Fig. 1c). We found a Flack factor of about 91%, which indicates that our samples are predominantly of a single structural chirality. Ab initio calculations of the electronic bulk band structure predict that both chiral crystals exhibit a threefold degeneracy at Γ near the Fermi level, EF (Fig. 1d, Extended Data Fig. 4), which is described by a low-energy Hamiltonian that exhibits a threefold fermion associated with Chern number +2 (refs 17,18). We refer to this Chern number as chiral charge—a usage of the term ‘chiral’ that is distinct from the notion of structural chirality defined above and that motivates our use of the term ‘chiral fermion’ to describe these topological band crossings. The R point hosts a fourfold degeneracy corresponding to a fourfold fermion with Chern number −2. These two higher-fold chiral fermions are pinned to opposite time-reversal-invariant momenta and are consequently constrained to be maximally separated in momentum space (Fig. 1e), suggesting that CoSi might provide a near-ideal platform for accessing topological phenomena using a variety of techniques. The hole pocket at M is topologically trivial at the band relevant for low-energy physics, but it is well separated in momentum space from the Γ and R topological crossings, and it is not expected to affect topological transport, such as the chiral anomaly (we note that the longitudinal conductivity σxx does not correspond to topological transport). The two higher-fold chiral fermions lead to a net Chern number of zero in the entire bulk Brillouin zone (BZ), as expected from theoretical considerations. CoSi and RhSi also satisfy a key criterion for an ideal topological conductor, namely, they have only two chiral fermions in the bulk BZ, the minimum non-zero number allowed6.

Fig. 1: Structural chirality and topological quantum chirality in CoSi and RhSi.
Fig. 1

a, Chiral crystal structure of XSi (X = Co, Rh), space group P213, number 198. b, Single-crystal X-ray diffraction precession pattern of the (0kl) planes of CoSi at 100 K. The resolved spots confirm space group P213 with lattice constant a = 4.43 Å. c, Three-dimensional Fourier map showing the electron density in the CoSi structure. The colour indicates electron density, with yellow corresponding to electron-rich regions and white to electron-poor regions. d, Ab initio calculation of the electronic bulk band structure along high-symmetry lines. A threefold-degenerate topological chiral fermion is predicted at Γ and a fourfold topological chiral fermion at R, with Chern numbers +2 and −2, respectively. The highest valence (blue) and lowest conduction (red) bands fix a topologically non-trivial energy window (green dashed lines). e, Bulk BZ and (001) surface BZ with high-symmetry points; the predicted chiral fermions (red and blue spheres) are marked. f, Ab initio calculation of the surface spectral weight on the (001) surface for CoSi, with the (001) surface BZ marked (red box). The predicted bulk chiral fermions project onto \(\bar{{\rm{\Gamma }}}\) and \(\bar{{\rm{M}}}\), connected by a pair of topological Fermi arcs extending diagonally across the surface BZ. The colour bar limits correspond to high (H) and low (L) electron density.

The two chiral fermions remain topologically non-trivial over a wide energy range. In particular, CoSi maintains constant-energy surfaces with non-zero Chern number over an energy window of 0.85 eV, whereas for RhSi this window is 1.3 eV (Fig. 1d). This prediction suggests that XSi satisfies another criterion for an ideal topological conductor: a large topologically non-trivial energy window. A Fermi surface calculated ab initio shows that the projection of the higher-fold chiral fermions to the (001) surface results in a hole (electron) pocket at \(\bar{{\rm{\Gamma }}}\left(\bar{{\rm{M}}}\right)\) with Chern number +2 (−2; Fig. 1f). As a result, we expect that XSi hosts Fermi arcs of length π spanning the entire surface BZ, again suggesting that these materials may realize a near-ideal topological conductor.

Using low-photon-energy ARPES, we experimentally study the (001) surface of CoSi and RhSi to reveal their surface electronic structure. For CoSi, the measured constant-energy contours show the following dominant features: two concentric contours around the \(\bar{{\rm{\Gamma }}}\) point, a faint contour at the \(\bar{{\rm{X}}}\) point and long winding states extending along the \(\bar{{\rm{M}}}-\bar{{\rm{\Gamma }}}-\bar{{\rm{M}}}\) direction (Fig. 2a). Both the \(\bar{{\rm{\Gamma }}}\) and \(\bar{{\rm{X}}}\) pockets show a hole-like behaviour (Extended Data Fig. 5). The measured surface electronic structure for RhSi shows similar features (Fig. 2b, Extended Data Fig. 1). Using only our spectra, we first sketch the key features of the experimental Fermi surface for CoSi (Fig. 2c). Then, to better understand the k-space trajectory of the long winding states in CoSi, we study Lorentzian fits to the momentum distribution curves (MDCs) of the ARPES spectrum (Extended Data Fig. 6). We plot the Lorentzian peak positions as the extracted band dispersion (Fig. 2d, e) and we find that the long winding states extend from the centre of the BZ to the \(\bar{{\rm{M}}}\) pocket (Fig. 2f). To better understand the nature of these states, we analyse the photon energy dependence of the ARPES results and find that the long winding states do not disperse as we vary the photon energy, suggesting that they are surface states (Extended Data Fig. 7). Moreover, we observe an overall agreement between the ARPES data and the Fermi surface calculated ab initio, where topological Fermi arcs connect the \(\bar{{\rm{\Gamma }}}\) and \(\bar{{\rm{M}}}\) pockets (Fig. 1f). Taken together, these results suggest that the long winding states observed in ARPES may be topological Fermi arcs.

Fig. 2: Fermiology of the (001) surface electronic topology in CoSi and RhSi.
Fig. 2

a, ARPES constant-energy contours for CoSi measured at incident photon energy of 50 eV, with the BZ boundary marked (rightmost panel, red dashed line). We observe long winding states connecting the \(\bar{{\rm{\Gamma }}}\) and \(\bar{{\rm{M}}}\) pockets. b, Fermi surface for RhSi measured at incident photon energy of 82 eV, with the BZ boundary marked (blue dashed line). Again, we observe long winding states extending diagonally across the BZ (Extended Data Fig. 1). c, Schematic of the measured Fermi surface for CoSi, showing topologically nontrivial hole- (h; green) and electron- (e; blue) like bulk pockets and long winding states (orange). The grey pocket is topologically trivial. d, Zoomed-in view of the long winding states, with a trajectory obtained by fitting Lorentzians to the MDCs of the ARPES spectrum (blue circles). e, Representative Lorentzian fit to the MDC along k1 for k2 = 0 at binding energy E − EF = −10 meV (Extended Data Fig. 6). f, Schematic overlaid with Lorentzian peaks extracted from the MDCs passing through the long winding states. We mark two straight cuts, I and II (magenta lines), as well as a closed-loop cut, \({\mathscr{P}}\) (magenta circle).

Grounded on the framework of topological band theory, the bulk–boundary correspondence of chiral fermions makes it possible to use ARPES (spectroscopic) measurements to determine the Chern numbers of a crystal by probing the surface-state dispersion (Fig. 3a, Methods). Such spectroscopic methods to determine Chern numbers have become well accepted in the field26. Using such an approach, we provide two spectral signatures of Fermi arcs in CoSi. We first look at the dispersion of the candidate Fermi arcs along a pair of energy–momentum cuts on opposite sides of the \(\bar{{\rm{\Gamma }}}\) pocket, taken at fixed momentum +kx (cut I) and −kx (cut II; Fig. 2f). In cut I, we observe two right-moving chiral edge modes (Fig. 3b, c). Because the cut passes through two BZs (Fig. 2f), we associate one right-moving mode with each BZ. Next, we fit Lorentzian peaks to the MDCs and we find that the extracted dispersion again suggests two chiral edge modes (Fig. 3d). Along cut II, we observe two left-moving chiral edge modes (Fig. 3e, f). Consequently, one chiral edge mode is observed for each measured surface BZ on cuts I and II, but with opposite Fermi velocity directions. In this way, our ARPES spectra suggest that the number of chiral edge modes n changes by +2 when the k-slice is swept from cut I to cut II. This again suggests that the long winding states are topological Fermi arcs. Moreover, these ARPES results imply that projected topological charge with net Chern number +2 resides near \(\bar{{\rm{\Gamma }}}\).

Fig. 3: Observation of topological chiral edge modes in CoSi.
Fig. 3

a, Left panel, schematic of three-dimensional bulk and two-dimensional surface BZ of CoSi with chiral fermions marked (±) and including several examples of two-dimensional manifolds hosting a Chern number n (green planes, red cylinder)33. Every plane in the bulk has a non-zero n. The cylinder enclosing the bulk chiral fermion at Γ has n = +2. Right panel, Fermi arcs (orange curves) show up as chiral edge modes (orange lines). Energy–momentum cuts (k||) on opposite sides of \(\bar{{\rm{\Gamma }}}\) are expected to show chiral modes propagating in opposite directions. b, ARPES spectrum along cut I (as marked in Fig. 2f), suggesting two right-moving chiral edge modes (black arrows). Vertical dashed lines mark the BZ boundaries. c, Second-derivative plot of cut I, with fitted Lorentzian peaks (blue dots) for a series of MDCs to track the dispersion. d, MDCs (blue circles) and Lorentzian fits (red lines) tracking the chiral edge modes (orange dashed lines) for cut I. e, Same as b, but for cut II, suggesting two left-moving chiral edge modes (black arrows). f, Same as c, but for cut II. The difference in the net number of chiral edge modes on cut I and cut II suggests a Chern number of +2 near \(\bar{{\rm{\Gamma }}}\).

Next we search for other Chern numbers encoded by the surface-state band structure. We study an ARPES energy–momentum cut on a loop \({\mathscr{P}}\) enclosing \(\bar{{\rm{M}}}\) (Fig. 4a, inset). Again following the bulk–boundary correspondence, we aim to extract the Chern number of chiral fermions projecting on \(\bar{{\rm{M}}}\). The cut \({\mathscr{P}}\) shows two right-moving chiral edge modes dispersing towards EF (Fig. 4a, b), suggesting a Chern number of −2 on the associated bulk manifold. Furthermore, the surface spectral weight along \({\mathscr{P}}\) calculated ab initio is consistent with our experimental results (Fig. 4c). Our ARPES spectra on cuts I, II and \({\mathscr{P}}\) suggest that CoSi hosts a projected chiral charge of +2 at \(\bar{{\rm{\Gamma }}}\) with its partner chiral charge of −2 projecting on \(\bar{{\rm{M}}}\). This again provides evidence that the long winding states are a pair of topological Fermi arcs that traverse the surface BZ on a diagonal, connecting the \(\bar{{\rm{\Gamma }}}\) and \(\bar{{\rm{M}}}\) pockets. Our ARPES spectra of RhSi also provide evidence for gigantic topological Fermi arcs following a similar analysis (Extended Data Fig. 1).

Fig. 4: Observation of helicoid quantum arcs in CoSi.
Fig. 4

a, ARPES spectrum along loop \({\mathscr{P}}\), revealing two right-moving chiral edge modes and a projected chiral charge of −2 at \(\bar{{\rm{M}}}\). Inset, definition of loop \({\mathscr{P}}\), starting from the green mark and proceeding clockwise. b, Lorentzian fits (red lines) to a series of MDCs (blue circles) along \({\mathscr{P}}\) to track the dispersion of the chiral edge modes (orange dashed lines). c, Ab initio calculation of the dispersion along a loop around \(\bar{{\rm{M}}}\), showing two right-moving chiral edge modes, consistent with the ARPES data. Colour scale as in Fig. 1f. d, Extracted dispersion of the chiral edge modes on \({\mathscr{P}}\) and a second inner loop from Lorentzian fits to the MDCs. Error bars correspond to the momentum resolution. Inset, definition of the second inner loop (black). We observe that the chiral edge modes spiral clockwise with decreasing binding energy (approaching EF). e, Perspective plot of d, where the two loops are shown as two concentric cylinders. The winding of the chiral modes around \(\bar{{\rm{M}}}\) as a function of binding energy suggests that the Fermi arcs have a helicoid structure6,27. f, Constant-energy contours calculated ab initio, consistent with the helicoid Fermi arc structure observed in our ARPES spectra.

To further explore the topological properties of CoSi, we examine in greater detail the structure of the Fermi arcs near \(\bar{{\rm{M}}}\). We consider the dispersion on \({\mathscr{P}}\) (plotted as a magenta loop in Fig. 4d, inset) and we also extract the dispersion from Lorentzian fitting on a second, tighter circle (black loop; Extended Data Fig. 8). We observe that as we decrease the binding energy (approaching EF), the extracted dispersion spirals clockwise on both loops, suggesting that as a given k point traverses the loop, the energy of the state does not return to its initial value after a full cycle. Such a non-trivial electronic dispersion directly signals a projected chiral charge at \(\bar{{\rm{M}}}\) (Fig. 4d). In fact, the extracted dispersion is characteristic of the helicoid structure of topological Fermi arcs as they wind around a chiral fermion (Fig. 4e), suggesting that CoSi and RhSi17 are rare examples of non-compact surfaces in nature6,27.

To further understand these experimental results, we consider the spectral weight for the (001) surface calculated ab initio, and we observe a pair of Fermi arcs winding around the \(\bar{{\rm{\Gamma }}}\) and \(\bar{{\rm{M}}}\) pockets counter-clockwise and clockwise, respectively, with decreasing binding energy (approaching EF; Fig. 4f). The clockwise winding around \(\bar{{\rm{M}}}\) is consistent with our ARPES observation of a −2 projected chiral charge. Moreover, from our ab initio calculations, we predict that the −2 charge projecting to \(\bar{{\rm{M}}}\) arises from a fourfold chiral fermion at the bulk R point (Fig. 1d). The +2 chiral charge, which we associate with \(\bar{{\rm{\Gamma }}}\) from ARPES (Fig. 3), is further consistent with the threefold chiral fermion predicted at the bulk Γ point. By fully accounting for the predicted topological charges in the experiment, our ARPES results suggest the manifestation of a topological chiral crystal in CoSi. We can similarly account for the predicted topological charges of multifold fermions in RhSi (Extended Data Figs. 1, 2).

The surface-state dispersions in our ARPES spectra, taken together with the topological bulk–boundary correspondence, demonstrate that CoSi and RhSi are topological chiral crystals6,7. This experimental result is further consistent with the numerical result of first-principles calculations of the surface-state dispersions and topological invariants. Unlike previously reported Weyl semimetals, the Fermi arcs that we observe in CoSi and RhSi stretch diagonally across the entire (001) surface BZ, from \(\bar{{\rm{\Gamma }}}\) to \(\bar{{\rm{M}}}\). In fact, the Fermi arcs in XSi are longer than those found in TaAs by a factor of thirty. Our surface band structure measurements also demonstrate two well-separated Fermi pockets carrying Chern number ±2. Lastly, for the first time to our knowledge, we observe in an electronic material the helicoid structure of topological Fermi arcs. Our results suggest that CoSi and RhSi are excellent candidates for studying topological phenomena distinct to chiral fermions by using a variety of techniques4,8,11.

Crucially for applications, the topologically non-trivial energy window in CoSi and RhSi is an order of magnitude larger than that in TaAs6,21,22, rendering its quantum properties robust against changes in surface chemical potential and disorder. Moreover, the energy offset between the higher-fold chiral fermions at Γ and R is predicted to be about 225 meV. Such an energy offset is essential for inducing the chiral magnetic effect and its optical analogue28, as well as the photogalvanic effect (optical)6,17,29, and our results on RhSi samples suggest that it can be used to observe a fourfold quantized circular photogalvanic current17. When coupled to a compatible superconductor, CoSi and RhSi are compelling platforms for studying the superconducting pairing of Fermi surfaces with non-zero Chern numbers, which may be promising for realizing a new type of topological superconducting phase proposed recently17,30, which can be probed with scanning tunnelling microscopy-based spectroscopy31. CoSi and RhSi further open the door to the exploration of other exotic quantum phenomena when combined with the isochemical material FeSi. Fe1−xCoxSi may simultaneously host k-space (chiral fermions) and real-space (skyrmions) topological defects and their interplay, which can also be probed by scanning tunnelling microscopy/spectroscopy31. Through our observation of a helicoid surface state and its topological properties, our results suggest the experimental discovery of the first structurally chiral crystal that is also topological. In this way, our work provides a next-generation platform that could be used in the further study of, and the search for, new types of topological chiral crystal6.

While the present article was under review, a related work studying bulk dispersion in CoSi was reported32.


Observation of topological Fermi arcs in RhSi

Motivated by an interest in establishing families of Kramers–Weyl and closely related topological materials9,34,35,36,37,38,39,40,41,42,43,44,45, we expand our experiment to include rhodium silicide, RhSi, an isoelectronic cousin of CoSi. RhSi crystallizes in a chiral cubic lattice, space group P213, number 198. The calculated electronic band structure is generally similar to that of CoSi (side-by-side comparison in Fig. 1d). Constant-energy contours measured by ARPES further suggest features similar to those of CoSi (Extended Data Fig. 1a). In particular, we observe contours at the \(\bar{{\rm{\Gamma }}}\), \(\bar{{\rm{X}}}\) and \(\bar{{\rm{Y}}}\) points and long winding states that extend diagonally from \(\bar{{\rm{\Gamma }}}\) to \(\bar{{\rm{M}}}\). Again taking advantage of the bulk–boundary correspondence, we focus on counting chiral edge modes to determine the topological nature of RhSi (Extended Data Fig. 1b, c). A second-derivative plot of the Fermi surface further suggests the presence of long states stretching diagonally across the BZ, motivating a study of energy–momentum cuts through these states (Extended Data Fig. 1d–g). The cuts show that the long states take the form of a right-moving chiral edge mode (cut I) and a left-moving chiral edge mode (cut II) on opposite sides of \(\bar{{\rm{\Gamma }}}\). The net difference in the number of right-moving chiral edge modes suggests that a Chern number of +2 projects to \(\bar{{\rm{\Gamma }}}\) and that the long states are topological Fermi arcs. Proceeding again by analogy to CoSi, we study the band structure along a loop \({\mathscr{M}}\) enclosing \(\bar{{\rm{M}}}\), and we observe two right-moving chiral edge modes, suggesting a Chern number of −2 at R. These results suggest that RhSi, an isoelectronic cousin of CoSi, is another example of a near-ideal topological conductor.

Helicoid structure of the arcs and topology in RhSi

We study the Fermi arcs near \(\bar{{\rm{M}}}\) in RhSi, by analogy with the analysis performed for CoSi (Fig. 4). First, we consider ARPES energy–momentum cuts on an inner and outer loop enclosing \(\bar{{\rm{M}}}\), and we observe signatures of two right-moving chiral edge modes on each loop, suggesting an enclosed projected chiral charge of −2 (Extended Data Fig. 2a–d). We further fit Lorentzians to the MDCs and we find that the extracted dispersion exhibits clockwise winding on both loops for decreasing binding energy (approaching EF; Extended Data Fig. 2e, f). This dispersion again suggests a chiral charge −2 at \(\bar{{\rm{M}}}\) and is characteristic of the helicoid structure of Fermi arcs as they wind around a chiral fermion. Lastly, we perform ab initio calculations of the constant-energy contours (Extended Data Fig. 2g), which exhibit the same winding pattern. These ARPES results suggest that RhSi, like CoSi, exhibits helicoid topological Fermi arcs.

Bulk–boundary correspondence in CoSi and RhSi

By measuring the surface-state band structure of a crystal, ARPES is capable of demonstrating Fermi arcs and counting Chern numbers33. We briefly review the details of this method. First, recall that topological invariants are typically defined for a gapped system. A topological conductor is by definition gapless so we cannot define a topological invariant for the full bulk BZ of the three-dimensional system. However, we may be able to choose certain two-dimensional k-space manifolds where the band structure is fully gapped, and a two-dimensional invariant can then be defined on this slice of momentum space (Fig. 3a). We consider specifically the case of a chiral fermion in a topological conductor and we study a gapped momentum-space slice that cuts in between two chiral fermions. By definition, a chiral fermion is associated with a non-zero Chern number, so at least some of these slices will be characterized by a non-zero Chern number. Following the bulk–boundary correspondence, these slices will then contribute chiral edge modes to the surface-state dispersion. If we image by collecting together the chiral edge modes from all of the gapped slices, we will assemble the entire topological Fermi arc of the topological conductor. If we run the bulk–boundary correspondence in reverse, we can instead measure the surface states by ARPES and count the chiral edge modes on a particular one-dimensional slice of the surface BZ to determine the Chern number of the underlying two-dimensional slice of the bulk BZ. These Chern numbers in turn fix the chiral charges of the topological fermions.

Next we highlight two spectral signatures that can determine the chiral charges specifically in XSi (X = Co, Rh). First, we consider chiral edge modes along straight k-slices. Theoretically, the two-dimensional k-slices on the two sides of the chiral fermion at Γ are related by time-reversal symmetry and therefore should have equal and opposite Chern numbers (nl = −nr; Fig. 3b). On the other hand, because the threefold chiral fermion at Γ is predicted to carry chiral charge +2, we expect that the difference should be nl − nr = +2, resulting in nl(r) = ±1. Therefore we expect one net left-moving chiral edge mode on one cut and one net right-moving chiral edge mode on the other. For the second signature, we study the chiral edge modes along a closed loop in the surface BZ. Any loop that encloses the projected chiral charge of +2 at \(\bar{{\rm{\Gamma }}}\) or −2 at \(\bar{{\rm{M}}}\) has this number of net chiral edge modes along this path. By taking advantage of this correspondence, we can count surface states on loops in our ARPES spectra to determine enclosed projected chiral charges.

Growth of CoSi and RhSi single crystals

Single crystals of CoSi were grown using a chemical vapour transport technique. First, polycrystalline CoSi was prepared by arc-melting stoichiometric amounts of Co slices and Si pieces, which were crushed and ground into a powder. The sample was then sealed in an evacuated silica tube with an iodine concentration of approximately 0.25 mg cm−3. The transport reaction took place at a temperature gradient from 1,000 °C (source) to 1,100 °C (sink) for two weeks. The resulting CoSi single crystals had a metallic lustre and varied in size from 1 mm to 2 mm. Single crystals of RhSi were grown from a melt using the vertical Bridgman crystal growth technique at a non-stoichiometric composition. In particular, we induced a slight excess of Si to ensure flux growth inside the Bridgman ampoule. First, a polycrystalline ingot was prepared by pre-melting the highly pure constituents under an argon atmosphere using an arc furnace. The crushed powder was poured into a custom-designed sharp-edged alumina tube and then sealed inside a tantalum tube under an argon atmosphere. The sample was heated to 1,550 °C and then slowly pooled to the cold zone at a rate of 0.8 mm h−1. Single crystals on average about 15 mm in length and about 6 mm in diameter were obtained.

X-ray diffraction

Single crystals of CoSi were mounted on the tips of Kapton loops. The low-temperature (100 K) intensity data were collected on a Bruker Apex II X-ray diffractometer with Mo radiation Kα1 (λ = 0.71073 Å−1). Measurements were performed over a full sphere in k-space with 0.5° scans in angle ω with an exposure time of 10 s per frame (Extended Data Fig. 3). The SMART software was used for data acquisition. The extracted intensities were corrected for Lorentz and polarization effects with the SAINT program. Numerical absorption corrections were accomplished with XPREP, which is based on face-indexed absorption46. The twin unit cell was tested. With the SHELXTL package, the crystal structures were solved using direct methods and refined by full-matrix least-squares on F2 (ref. 47) No vacancies were observed according to the refinement and no residual electron density was detected, indicating that the CoSi crystals were of high quality.

Sample surface preparation for ARPES

For CoSi single crystals, the surface preparation procedure followed the conventional in situ mechanical cleaving approach. This cleaving method resulted in a low success rate and the resulting surface typically appeared rough under a microscope. We speculate that the difficulty in cleaving CoSi single crystals may be a result of its cubic structure, strong covalent bonding and lack of a preferred cleaving plane. For RhSi single crystals, their large size enabled mechanical cutting and polishing along the (001) surface. An in situ sputtering and annealing procedure, combined with low-energy/reflection high-energy electron diffraction characterization, was used to obtain a clean surface suitable for ARPES measurements. The typical spectral line width was narrowed and the background signal was reduced for these samples compared with the mechanically cleaved CoSi samples.


ARPES measurements were carried out at beamlines 10.0.1 and 4.0.3 at the Advanced Light Source in Berkeley, California, USA. A Scienta R4000 electron analyser was used at beamline10.0.1 and a Scienta R8000 was used at beamline 4.0.3. At both beamlines the angular resolution was <0.2° and the energy resolution was better than 20 meV. Samples were cleaved or sputtered/annealed in situ and measured under vacuum better than 5 × 10−11 torr at T < 20 K.

First-principles calculations

Numerical calculations of XSi (X = Co, Rh) were performed within the density functional theory framework using the OPENMX package and the full potential augmented plane-wave method as implemented in the package WIEN2k48,49,50. The generalized gradient approximation was used51. Experimentally measured lattice constants were used in the density functional theory calculations of material band structures52. A Γ-centred k-point 10 × 10 × 10 mesh was used and spin–orbit coupling was included in self-consistent cycles. To generate the (001) surface states of CoSi and RhSi, Wannier functions were generated using the p orbitals of Si and the d orbitals of Co and Rh. The surface states were calculated for a semi-infinite slab by the iterative Green’s function method.

Electronic bulk band structure

The electronic bulk band structure of CoSi is shown with and without spin–orbit coupling in Extended Data Fig. 4. The small energy splitting that arises from introducing spin–orbit coupling is approximately 40 meV, which is negligible compared to the about 1.2 eV topologically non-trivial energy window. Our results suggest that we do not need to consider spin–orbit coupling, either theoretically or experimentally, to demonstrate a chiral charge in CoSi.

Tracking the Fermi arcs by fitting ARPES MDCs

The following procedure was performed to track the Fermi arcs in the surface BZ of CoSi (Extended Data Fig. 6a). At E = EF, MDCs were collected for various fixed ky values (extending along the region of interest). Each MDC was fitted with a Lorentzian function to pinpoint the kx value that corresponds to the peak maximum. The peak maximum of the fitted MDC is plotted on top of the measured Fermi surface and marked with blue circles in Extended Data Fig. 6b. Where necessary, the peak maximum corresponding to the \(\bar{{\rm{X}}}\) pockets is annotated on each MDC. For the chiral edge modes discussed in the main text and shown in the Extended Data, a similar MDC fitting procedure was used at different binding energies to track the dispersion.

Study of Fermi arcs with varying photon energy

Extended Data Fig. 7 shows an MDC cutting through the Fermi arc as a function of incident photon energy. Using methods established previously8 we find that within the experimental resolution the Fermi arc shows negligible variation from 80 eV to 110 eV, providing additional evidence that it is indeed a surface state.

Data availability

The data supporting the findings of this study are available within the paper, and other findings of this study are available from the corresponding author upon reasonable request.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.


  1. 1.

    Castelvecchi, D. The strange topology that is reshaping physics. Nature 547, 272–274 (2017).

  2. 2.

    Hasan, M. Z. & Kane, C. L. Topological insulators. Rev. Mod. Phys. 82, 3045–3067 (2010).

  3. 3.

    Qi, X. L. & Zhang, S. C. Topological insulators and superconductors. Rev. Mod. Phys. 83, 1057–1110 (2011).

  4. 4.

    Vafek, O. & Vishwanath, A. From high-T c cuprates and graphene to topological insulators and Weyl semimetals. Annu. Rev. Condens. Matter Phys. 5, 83–112 (2014).

  5. 5.

    Hasan, M. et al. Topological insulators, topological superconductors and Weyl semimetals. Phys. Scr. 2015, 014001 (2015); corrigendum 2016, 019501 (2016).

  6. 6.

    Chang, G. et al. Topological quantum properties of chiral crystals. Nat. Mater. 17, 978–985 (2018).

  7. 7.

    Shekhar, C. Chirality meets topology. Nat. Mater. 17, 953–954 (2018).

  8. 8.

    Hasan, M. Z. et al. Discovery of Weyl fermion semimetals and topological Fermi arcs. Annu. Rev. Condens. Matter Phys. 8, 289–309 (2017).

  9. 9.

    Jia, S. et al. Weyl semimetals, Fermi arcs and chiral anomaly. Nat. Mater. 15, 1140–1144 (2016).

  10. 10.

    Zhang, C. et al. Signatures of the Adler–Bell–Jackiw chiral anomaly in a Weyl semimetal. Nat. Commun. 7, 10735 (2016).

  11. 11.

    Yan, B. & Felser, C. Topological materials. Annu. Rev. Condens. Matter Phys. 8, 337–354 (2017).

  12. 12.

    Murakami, S. Phase transition between the quantum spin hall and insulator phases in 3D: emergence of a topological gapless phase. New J. Phys. 9, 356 (2007); corrigendum 10, 029802 (2008).

  13. 13.

    Wan, X. et al. Topological semimetal and Fermi-arc surface states in the electronic structure of pyrochlore iridates. Phys. Rev. B 83, 205101 (2011).

  14. 14.

    Burkov, A. A. & Balents, L. Weyl semimetal in a topological insulator multilayer. Phys. Rev. Lett. 107, 127205 (2011).

  15. 15.

    Xu, G. et al. Chern semimetal and the quantized anomalous Hall effect in HgCr2Se4. Phys. Rev. Lett. 107, 186806 (2011).

  16. 16.

    Bradlyn, B. et al. Beyond Dirac and Weyl fermions: unconventional quasiparticles in conventional crystals. Science 353, aaf5037 (2016).

  17. 17.

    Chang, G. et al. Unconventional chiral fermions and large topological Fermi arcs in RhSi. Phys. Rev. Lett. 119, 206401 (2017).

  18. 18.

    Tang, P. et al. Multiple types of topological fermions in transition metal silicides. Phys. Rev. Lett. 119, 206402 (2017).

  19. 19.

    Watanabe, H. et al. Filling constraints for spin–orbit coupled insulators in symmorphic and nonsymmorphic crystals. Proc. Natl Acad. Sci. USA 112, 14551–14556 (2015).

  20. 20.

    Huang, S.-M. et al. A Weyl fermion semimetal with surface Fermi arcs in the transition metal monopnictide TaAs class. Nat. Commun. 6, 7373 (2015).

  21. 21.

    Xu, S.-Y. et al. Discovery of a Weyl fermion semimetal and topological Fermi arcs. Science 349, 613–617 (2015).

  22. 22.

    Lv, B. Q. et al. Experimental discovery of Weyl semimetal TaAs. Phys. Rev. X 5, 031013 (2015).

  23. 23.

    Weng, H. et al. Weyl semimetal phase in non-centrosymmetric transition metal monophosphides. Phys. Rev. X 5, 011029 (2015).

  24. 24.

    Wieder, B. J. et al. Spin-orbit semimetals in the layer groups. Phys. Rev. B 94, 155108 (2016).

  25. 25.

    Mañes, J. L. Existence of bulk chiral fermions and crystal symmetry. Phys. Rev. B 85, 155118 (2012).

  26. 26.

    Royal Swedish Academy of Sciences. Topological phase transitions and topological phases of matter. Nobel Media AB https://www.nobelprize.org/uploads/2018/06/advanced-physicsprize2016-1.pdf (2016).

  27. 27.

    Fang, C. et al. Topological semimetals with helicoid surface states. Nat. Phys. 12, 936–941 (2016).

  28. 28.

    Zyuzin, A. A. et al. Weyl semimetal with broken time reversal and inversion symmetries. Phys. Rev. B 85, 165110 (2012).

  29. 29.

    de Juan, F. et al. Quantized circular photogalvanic effect in Weyl semimetals. Nat. Commun. 8, 15995 (2017).

  30. 30.

    Li, Y. & Haldane, F. D. M. Topological nodal Cooper pairing in doped Weyl metals. Phys. Rev. B 120, 067003 (2018).

  31. 31.

    Yin, J.-X. et al. Giant and anisotropic many-body spin–orbit tunability in a strongly correlated kagome magnet. Nature 562, 91–95 (2018).

  32. 32.

    Takane, D. et al. Observation of chiral fermions with a large topological charge and associated Fermi-arc surface states in CoSi. Preprint at https://arxiv.org/abs/1809.01312 (2018).

  33. 33.

    Belopolski, I. et al. Criteria for directly detecting topological Fermi arcs in Weyl semimetals. Phys. Rev. Lett. 116, 066802 (2016).

  34. 34.

    Chang, G. et al. Kramers theorem-enforced Weyl fermions: theory and materials predictions (Ag3BO3, TlTe2O6 and Ag2Se related families). Preprint at https://arxiv.org/abs/1611.07925v1 (2016).

  35. 35.

    Chang, G. et al. Universal topological electronic properties of nonmagnetic chiral crystals. In APS March Meeting 2018 https://meetings.aps.org/Meeting/MAR18/Session/Y14.14 (APS, 2018).

  36. 36.

    Zhang, C.-L. et al. Ultraquantum magnetoresistance in the Kramers–Weyl semimetal candidate β-Ag2Se. Phys. Rev. B 96, 165148 (2017).

  37. 37.

    Keimer, B. & Moore, J. E. The physics of quantum materials. Nat. Phys. 13, 1045–1055 (2017).

  38. 38.

    Burkov, A. A. Weyl metals. Annu. Rev. Condens. Matter Phys. 9, 359–378 (2018).

  39. 39.

    Weyl, H. Elektron und gravitation. I. Z. Phys. 56, 330–352 (1929).

  40. 40.

    Nielsen, H. B. & Ninomiya, M. The Adler–Bell–Jackiw anomaly and Weyl fermions in a crystal. Phys. Lett. B 130, 389–396 (1983).

  41. 41.

    Soluyanov, A. A. et al. Type-II Weyl semimetals. Nature 527, 495–498 (2015).

  42. 42.

    Huang, L. et al. Spectroscopic evidence for a type II Weyl semimetallic state in MoTe2. Nat. Mater. 15, 1155–1160 (2016).

  43. 43.

    Deng, K. et al. Experimental observation of topological Fermi arcs in type-II Weyl semimetal MoTe2. Nat. Phys. 12, 1105–1110 (2016).

  44. 44.

    Wu, Y. et al. Observation of Fermi arcs in the type-II Weyl semimetal candidate WTe2. Phys. Rev. B 94, 121113 (2016).

  45. 45.

    Belopolski, I. et al. Fermi arc electronic structure and Chern numbers in MoxW1−xTe2. Phys. Rev. B 94, 085127 (2016).

  46. 46.

    Rastogi, R., Srivastava, S. K., Asolia, S. & Butcher, R. J. Synthesis, spectral characterization and ligation of N-[2-(phenylseleno)ethyl]phthalimide. J. Cryst. Process Technol. 3, 31–35 (2013).

  47. 47.

    Sheldrick, G. M. A short history of SHELX. Acta Crystallogr. A 64, 112–122 (2008).

  48. 48.

    Ozaki, T. et al. OpenMX http://www.openmx-square.org/.

  49. 49.

    Kresse, G. & Furthmueller, G. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B 54, 11169–11186 (1996).

  50. 50.

    Kresse, G. & Joubert, D. From ultrasoft pseudopotentials to the projector augmented-wave method. Phys. Rev. B 59, 1758–1775 (1999).

  51. 51.

    Perdew, J. P. et al. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865–3868 (1996).

  52. 52.

    Demchenko, P. et al. Single crystal investigation of the new phase Er0.85Co4.31Si and of CoSi. Chem. Met. Alloys 1, 50–53 (2008).

Download references


Work at Princeton and Princeton-led ARPES measurements and theoretical work reported here were supported by the US Department of Energy under a Basic Energy Sciences grant (number DOE/BES DE-FG-02-05ER46200). M.Z.H. acknowledges a Visiting Scientistship at the Materials Science Division of Lawrence Berkeley National Laboratory. This research used resources of the Advanced Light Source, which is a DOE Office of Science Facility, under contract number DE-AC02-05CH11231. S.J. was supported by the National Natural Science Foundation of China (U1832214, 11774007), the National Key R&D Program of China (2018YFA0305601) and the Key Research Program of the Chinese Academy of Science (grant number XDPB08-1). H.L. acknowledges Academia Sinica, Taiwan for support under the Innovative Materials and Analysis Technology Exploration (AS-iMATE-107-11) programme. T.-R.C. is supported by the Young Scholar Fellowship Program of the Ministry of Science and Technology (MOST) in Taiwan, under the MOST Grant for the Columbus Program (MOST107-2636-M-006-004), National Cheng Kung University, Taiwan, and National Center for Theoretical Sciences (NCTS), Taiwan. C.F. acknowledges financial support by the ERC through Advanced Grant number 291472 (Idea Heusler) and 742068 (TOPMAT). The authors thank S. K. Mo and J. D. Denlinger for beamline support at the Advanced Light Source in Berkeley, California. We thank R. Dhall of the Molecular Foundry at LBL in Berkeley, California for assistance. M.Z.H. acknowledges support from the Miller Institute of Basic Research in Science at the University of California in Berkeley in the form of a Visiting Miller Professorship during the early stages of this work. We also thank G. Bian and H. Verlinde for discussions.

Author information

Author notes

  1. These authors contributed equally: Daniel S. Sanchez, Ilya Belopolski, Tyler A. Cochran


  1. Laboratory for Topological Quantum Matter and Advanced Spectroscopy (B7), Department of Physics, Princeton University, Princeton, NJ, USA

    • Daniel S. Sanchez
    • , Ilya Belopolski
    • , Tyler A. Cochran
    • , Jia-Xin Yin
    • , Guoqing Chang
    • , Nasser Alidoust
    • , Daniel Multer
    • , Songtian S. Zhang
    • , Nana Shumiya
    • , Su-Yang Xu
    •  & M. Zahid Hasan
  2. International Center for Quantum Materials, School of Physics, Peking University, Beijing, China

    • Xitong Xu
    • , Xirui Wang
    • , Guang-Qiang Wang
    •  & Shuang Jia
  3. Department of Chemistry, Louisiana State University, Baton Rouge, LA, USA

    • Weiwei Xie
  4. Max Planck Institute for Chemical Physics of Solids, Dresden, Germany

    • Kaustuv Manna
    • , Vicky Süß
    •  & Claudia Felser
  5. Institute of Physics, Academia Sinica, Taipei, Taiwan

    • Cheng-Yi Huang
    •  & Hsin Lin
  6. Rigetti Quantum Computing, Berkeley, CA, USA

    • Nasser Alidoust
  7. Department of Chemistry, Princeton University, Princeton, NJ, USA

    • Daniel Multer
  8. Department of Physics, National Cheng Kung University, Tainan, Taiwan

    • Tay-Rong Chang
  9. Collaborative Innovation Center of Quantum Matter, Beijing, China

    • Shuang Jia
  10. CAS Center for Excellence in Topological Quantum Computation, University of the Chinese Academy of Science, Beijing, China

    • Shuang Jia
  11. Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA, USA

    • M. Zahid Hasan


  1. Search for Daniel S. Sanchez in:

  2. Search for Ilya Belopolski in:

  3. Search for Tyler A. Cochran in:

  4. Search for Xitong Xu in:

  5. Search for Jia-Xin Yin in:

  6. Search for Guoqing Chang in:

  7. Search for Weiwei Xie in:

  8. Search for Kaustuv Manna in:

  9. Search for Vicky Süß in:

  10. Search for Cheng-Yi Huang in:

  11. Search for Nasser Alidoust in:

  12. Search for Daniel Multer in:

  13. Search for Songtian S. Zhang in:

  14. Search for Nana Shumiya in:

  15. Search for Xirui Wang in:

  16. Search for Guang-Qiang Wang in:

  17. Search for Tay-Rong Chang in:

  18. Search for Claudia Felser in:

  19. Search for Su-Yang Xu in:

  20. Search for Shuang Jia in:

  21. Search for Hsin Lin in:

  22. Search for M. Zahid Hasan in:


D.S.S. and M.Z.H. conceived the project. D.S.S., I.B., T.A.C. and N.A. conducted the ARPES experiments in consultation with S.-Y.X. and M.Z.H.; X.X., W.X., X.W., G.-Q.W. and S.J. synthesized and characterized the RhSi and CoSi samples; K.M., V.S. and C.F. synthesized the RhSi samples; G.C., C.-Y.H., T.-R.C. and H.L. performed the first-principles/density functional theory calculations; D.S.S., I.B. and T.A.C. performed the analysis, interpretation and figure development in consultation with J.-X.Y., G.C., N.A., D.M., S.S.Z., N.S., S.-Y.X. and M.Z.H.; D.S.S. wrote the manuscript in consultation with I.B., J.-X.Y., G.C., S.-Y.X. and M.Z.H.; M.Z.H. supervised the project.

Competing interests

The authors declare no competing interests.

Corresponding author

Correspondence to M. Zahid Hasan.

Extended data figures and tables

  1. Extended Data Fig. 1 Topological arcs in the structurally chiral crystal RhSi.

    a, ARPES-measured Fermi surface and constant-binding-energy contours with an incident photon energy of 82 eV at 10 K. The BZ boundary is marked in blue. b, Three-dimensional bulk and two-dimensional surface BZ with higher-fold chiral fermions (±). The planes outlined in blue and the red cylinder are two-dimensional manifolds with indicated Chern number33 n. The cylinder enclosing the bulk chiral fermion at Γ has n = +2. c, Fermi arcs (orange) connect the projected chiral fermions. Energy dispersion cuts show that these two chiral edge modes are time-reversed partners propagating in opposite directions. d, Second-derivative Fermi surface with the straight and loop cuts of interest marked. e, ARPES spectrum along a loop \({\mathscr{M}}\), showing two right-moving chiral edge modes, suggesting that the fourfold chiral fermion at R carries Chern number −2. f, ARPES spectrum along cut I, showing a right-moving chiral edge mode. g, ARPES spectrum along cut II on the opposite side of the threefold chiral fermion at \(\bar{{\rm{\Gamma }}}\) (illustrated by the blue sphere), showing a left-moving chiral edge mode.

  2. Extended Data Fig. 2 Helicoid arcs in RhSi.

    a, Energy dispersion cut on an inner loop of radius 0.18π/a enclosing \(\bar{{\rm{M}}}\). b, Lorentzian fits (red traces) to the MDCs (blue dots) to track the observed chiral edge modes. c, d, Similar analysis to a and b, but for an outer loop of radius 0.23π/a. Black arrows show the Fermi velocity direction for the chiral edge modes. e, f, Top (e) and perspective (f) view of the helicoid dispersion extracted from the MDCs, plotted on two concentric loops, suggesting a clockwise spiral with decreasing binding energy. g, Constant-energy contours calculated ab initio, showing a consistent helicoid structure.

  3. Extended Data Fig. 3 X-ray diffraction study.

    Single-crystal X-ray diffraction precession image of the (0kl) planes in the reciprocal lattice of CoSi. The resolved spots from scans I and III are consistent with space group P213 at 100 K. This is also shown in the reflection intersection, scan II.

  4. Extended Data Fig. 4 Electronic bulk band topology.

    a, Band structure of CoSi in the absence of spin–orbit coupling interactions. b, Band structure in the presence of spin–orbit coupling. The highest valence and lowest conduction bands are coloured in blue and red, respectively.

  5. Extended Data Fig. 5 Surveying the \(\bar{{\rm{\Gamma }}}\) and \(\bar{{\rm{X}}}\) pockets on the (001) surface of CoSi.

    a, ARPES-measured Fermi surface with the BZ boundary (red dashed line) and the location of the energy–momentum cut of interest (magenta dashed line). b, \(\bar{{\rm{\Gamma }}}-\bar{{\rm{X}}}-\bar{{\rm{\Gamma }}}\) high-symmetry-line energy–momentum cut. c, Second-derivative plot of b. We observe an outer and an inner hole-like band at \(\bar{{\rm{\Gamma }}}\), whereas at \(\bar{{\rm{X}}}\) we observe signatures of a single hole-like band.

  6. Extended Data Fig. 6 Tracking the Fermi arcs by fitting Lorentzians to MDCs.

    a, Zoomed-in region of the Fermi surface (Fig. 2d), with the Fermi arcs tracked (blue circles) and the surface BZ marked (red dashed lines). b, Representative fits of Lorentzian functions (red lines) to the MDCs (filled blue circles). The peaks indicate the extracted positions of the Fermi arc (open blue circles) and the \(\bar{{\rm{X}}}\) pocket.

  7. Extended Data Fig. 7 Photon-energy dependence of the Fermi arc in CoSi.

    a, Location of the in-plane momentum direction (black arrow) along which the photon-energy dependence was studied, plotted on the Fermi surface (Fig. 2d) with the surface BZ (red dotted lines). b, MDCs at EF along the in-plane momentum direction illustrated in a, obtained for a series of photon energies from 80 eV to 110 eV in steps of 2 eV. The peak associated with the Fermi arc shows negligible variation in photon energy (dashed line) within the experimental resolution, providing further evidence that the observed Fermi arc is indeed a surface state.

  8. Extended Data Fig. 8 Systematics for the helicoid fitting in CoSi.

    a, Fermi arc trajectory extracted from Lorentzian fits to the MDCs (blue open circles) near \(\bar{{\rm{M}}}\), with overlaid schematic of the observed features (orange lines). There are two closed contours enclosing \(\bar{{\rm{M}}}\) on which we count chiral edge modes: the outer loop (magenta, Fig. 4) and the inner loop (black). b, Energy–momentum cut along the inner loop (radius 0.14π/a), starting from the green notch in a and winding clockwise (black arrow in a). c, Lorentzian fits (red curves) to the MDCs (blue circles) to extract the helicoid dispersion of the Fermi arcs. We observe two right-moving chiral edge modes dispersing towards EF (dashed orange lines). The corresponding bulk manifold has Chern number13 n = −2.

  9. Extended Data Table 1 Single-crystal crystallographic data for CoSi at 100(2) K and 300(2) K

About this article

Publication history




Issue Date



Further reading


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.