betaFeedback

WorksPhysics and Astronomy

Observation of light-induced topological state in graphene

Abstract

Driving quantum materials with light can reshape not only their electronic band structures but also the geometry of their electronic wavefunctions. In graphene, circularly polarized light is predicted to open gaps at the Dirac points and induce a pseudospin texture and Berry curvature characteristic of the Haldane model, realizing a Floquet Chern insulator. Transport and all-optical measurements have provided evidence for a light-induced anomalous Hall response, but cannot resolve the underlying electronic states in both energy and momentum space. Time- and angle-resolved photoemission spectroscopy has revealed Floquet-Bloch sidebands and hybridization gaps at crossings between Floquet bands. However, these experiments used linearly polarized driving, which preserves time-reversal symmetry and thus does not induce a topological phase. Here, we combine circular-dichroic time-resolved extreme-ultraviolet momentum microscopy with ab initio theory to investigate graphene driven by circularly polarized light. We observe a pump-induced photoemission circular dichroism near the Dirac point that reverses sign with pump helicity and across the Dirac energy, in agreement with our calculations. These results link the measured dichroic response to the light-induced pseudospin texture and its associated Berry curvature, providing spectroscopic evidence for the Floquet-engineered topological insulating state in graphene.

The field of Floquet-Bloch engineering [2, 11, 12], which exploits periodic optical driving to tailor a material’s electronic structure and topology, was propelled by Oka and Aoki’s pioneering proposal [1]. They predicted that circularly polarized light can induce a Haldane-like effective Hamiltonian [4] in monolayer graphene, giving rise to a Floquet topological insulator. In this scheme, the periodic drive breaks time-reversal symmetry and opens a gap at the Dirac points. Such Floquet topological states have since been experimentally realized in cold-atom optical lattices and photonic systems [13, 14].

In condensed matter, successful experimental demonstrations of Floquet physics have been largely limited to trivial and topological insulators [15, 16, 17, 18, 19, 20, 21, 22], and, very recently, a transient topological state attributed to a resonantly driven Floquet band inversion was reported in the semiconductor SnTe [23]. In particular, realizing and observing Floquet-Bloch states in graphene has proven challenging [24]. Graphene is a semimetal characterized by unusually fast electron-electron and electron-phonon scattering [25, 26, 27], raising the question of whether coherent Floquet dressing can coexist with such rapid scattering and decoherence processes [24, 28, 29]. Transport and all-optical measurements reported a light-induced anomalous Hall effect in graphene [5, 6]. However, because these measurements are integrated over both energy and momentum, they do not allow the topological Floquet–Bloch contribution to be disentangled from effects arising from asymmetric photocarrier distributions [30], leaving the microscopic origin of the observed phenomena only partially resolved. Direct spectroscopic evidence of light-induced Floquet topological states in graphene driven using circularly polarized light remains elusive and would require time- and angle-resolved photoemission spectroscopy (trARPES) measurements, which would enable the transient electronic structure to be probed as a function of both energy and momentum.

Very recent trARPES experiments have succeeded in revealing coherent Floquet-Bloch states in graphene under linearly polarized driving by resolving quantum-path interference between Floquet and Volkov sidebands [7, 8]. Floquet-induced hybridization gaps at the crossing between equilibrium and Floquet bands have subsequently been resolved [9]. However, this avoided-crossing gap is distinct from the Haldane-type mass gap predicted to open at the Dirac points. Indeed, these trARPES experiments [7, 8, 9] used linearly polarized driving, which preserves time-reversal symmetry and therefore does not open a Haldane mass at the Dirac points. The predicted Haldane-type gap, together with the associated pseudospin and Berry curvature around the Dirac points, thus remains to be investigated by trARPES.

Merely observing a gap, however, does not prove the existence of a Floquet topological insulator. Indeed, a gap can arise from many mechanisms unrelated to topology [31, 32, 33] and carries no information on the pseudospin texture and Berry curvature that define the topological phase. The same applies to other band-structure signatures, such as transient band inversions or light-induced Dirac cones [23]: they constrain the spectrum, but the topological character resides in the geometry of the light-dressed wave-functions. Recent theoretical work has shown that these quantities can be revealed through circular dichroism in trARPES [34]. Even when electron-electron and electron-phonon scattering broadens or obscures the Floquet-induced Haldane gap, the sign and momentum dependence of dichroic signals are predicted to retain information about the light-induced pseudospin [34]. A related approach was recently demonstrated on the surface of a three-dimensional topological insulator using ultraviolet circular-dichroic trARPES [35]. For graphene, where the Dirac cones lie at the Brillouin-zone boundary, it is necessary to use higher photon energies, which the recent developments of our circularly polarized extreme-ultraviolet (XUV) source make possible [10].

Here, we report measurements of pump-helicity-dependent photoemission circular dichroism in monolayer graphene using time- and polarization-resolved XUV momentum microscopy. Comparing the XUV dichroisms obtained with the two circular polarizations of the IR drive, we extract a double-dichroism that reverses sign across the Dirac point, in quantitative agreement with ab initio calculations. This behavior is consistent with the predicted light-induced pseudospin texture and its associated Berry curvature, finally providing direct spectroscopic evidence for a Floquet topological state in graphene.

Generation and detection of a light-induced Floquet topological insulator In equilibrium, graphene hosts massless Dirac fermions whose lattice pseudospin lies in-plane and points radially in or out of the Dirac cones (Figure 1a, left), with opposite chirality at K and K′. The wavefunction amplitudes on the two sublattices are equal, while their relative phase depends on momentum. In the high-frequency limit, circularly polarized light breaks time-reversal symmetry and generates an effective Haldane mass mFm_F in graphene [1, 36], opening a gap of 2∣mF∣2\lvert m_F\rvert at the Dirac points, which scales as mF∝E02/ωpump3m_F \propto E_0^2/\omega_{\mathrm{pump}}^3, where E0E_0 is the electric-field amplitude of the pump pulse and ωpump\omega_{\mathrm{pump}} its angular frequency. The sign of the mass is determined by the pump helicity [1, 34, 37].

Schematic of light-induced pseudospin, Berry curvature, and static photoemission circular dichroism in graphene

Figure 1. Light-induced pseudospin, Berry curvature, and static photoemission circular dichroism in graphene (a) Schematic of the Dirac bands and associated sublattice pseudospin (arrows) in equilibrium (left) and under circularly polarized IR driving (right). In equilibrium, the pseudospin lies in-plane (σz=0\sigma_z = 0) and the A and B sublattice weights are equal. The infrared (IR) driving pulse opens a gap, tilts the pseudospin out-of-plane (σz<0\sigma_z < 0 in blue, σz>0\sigma_z > 0 in red), imbalances the A and B sublattice weights (grey, red/blue spheres) and induces a Berry curvature Ω\Omega (red/blue colorbar) of opposite sign in the two bands. (b) Calculated CDAD for free-standing graphene without pump, with right- (σ+\sigma^+) and left- (σ−\sigma^-) circularly polarized pump pulses, and the resulting double-helicity dichroism (DD), for a drive strength of 4000 kV/cm. (c) Scheme of the experiment: a polarization-tunable infrared pump (IR, 1.2 eV, 135 fs) and a polarization-tunable XUV probe (21.6 eV) are focused on a graphene/SiC sample in the interaction chamber of a time-of-flight momentum microscope, at an incidence angle of 65°, with the light incidence plane along Γ\Gamma-K. (d) Measured normalized photoemission intensity (left) and CDAD (right) of graphene/SiC without pump, along the momentum direction perpendicular to Γ\Gamma-K through the Dirac point.

The out-of-plane component of the pseudospin, σz(k)=PA(k)−PB(k)\sigma_z(\mathbf{k}) = P_A(\mathbf{k}) - P_B(\mathbf{k}), is given by the difference between the probabilities PAP_A and PBP_B of finding the electron on the A and B sublattices, respectively, and vanishes in equilibrium (Figure 1a, left). Under circularly polarized driving, σz\sigma_z becomes finite, with opposite signs in the upper and lower bands. Reversing the pump helicity reverses its sign. The resulting pseudospin texture is associated with a finite Berry curvature (Figure 1a, right). This gives rise to a nonzero Chern number C=(1/2π)∫dk Ω(k)=1C = (1/2\pi) \int\mathrm{d}\mathbf{k}\,\Omega(\mathbf{k}) = 1, and an associated orbital magnetic moment Mz(k)=∓eℏ[ε+(k)−ε−(k)]Ω(k)M_z(\mathbf{k}) = \mp\frac{e}{\hbar}[\varepsilon_{+}(\mathbf{k}) - \varepsilon_{-}(\mathbf{k})]\Omega(\mathbf{k}), where ε±(k)\varepsilon_{\pm}(\mathbf{k}) is the upper/lower band, both of which are linked to the light-induced anomalous Hall response [5].

The pseudospin texture can be probed using circular dichroism in the angular distribution (CDAD) of photoelectrons emitted from the graphene π\pi bands. The photoemission intensity depends on interference between photoemission amplitudes from the two sublattices [38, 39, 40]. Under a circularly polarized pump, the leading term of the light-induced intrinsic photoemission XUV circular dichroism is proportional to σz(k)\sigma_z(\mathbf{k}) [34, 37, 41]. Consequently, CDAD can probe the light-induced sublattice imbalance and, through its relation to the Floquet Hamiltonian, provide information about the associated Berry curvature. As discussed below in connection with Figure 4, the pseudospin extends further away from the Dirac point than the Berry curvature, making the dichroic signal robust against scattering-induced broadening of the Floquet-induced Haldane gap.

Figure 4. Link between the light-induced pseudospin and the Berry curvature. (a,b) Band-resolved out-of-plane pseudospin expectation value σz\sigma_z (a) and Berry curvature Ω\Omega (b), overlaid on the Floquet band structure along kyk_y for a drive amplitude of 3000 kV/cm. (c) Full width at half maximum (FWHM) of σz\sigma_z and Ω\Omega as a function of the field strength. (d,e) Momentum profiles of σz\sigma_z (d) and Ω\Omega (e) for field strengths of 2000 and 4000 kV/cm. The inset in (e) shows the tails of Ω\Omega. (f) Induced band gap as a function of the field strength. Calculations use the Floquet Hamiltonian of a tight-binding model of graphene (see Methods). Gray areas in (c) and (d) highlight the experimental field strength.

The calculated CDAD of free-standing graphene at equilibrium and driven by a circular pump (for both helicities, σ+\sigma^{+} and σ−\sigma^{-}) are shown in Figure 1b (see Methods). The light-induced contribution is concentrated near the Dirac point, and changes sign with the pump helicity and across the Dirac point. To isolate this contribution, we define the double-helicity dichroism (DD) as DD=CDADσ−−CDADσ+DD = \mathrm{CDAD}_{\sigma^-} - \mathrm{CDAD}_{\sigma^+}, following an approach recently used for investigating Floquet-Bloch states on the surface of a three-dimensional topological insulator [35]. This doubly differential observable removes the equilibrium dichroism and other pump-helicity-independent contributions, isolating the helicity-odd response. The resulting DD is expected to be strongest near the Dirac point and to change sign across the Dirac energy (Figure 1b, right).

Experimental realization of light-induced topological state. A schematic of our experimental setup is shown in Figure 1c. In a nutshell, a polarization-tunable infrared pump pulse (1.2 eV, 135 fs, incident fluence 3.5 mJ/cm²) drives monolayer graphene, while a polarization-tunable XUV probe pulse (21.6 eV) ejects photoelectrons. The measurements are performed with the two pulses overlapped in time. The photoelectrons are collected using a time-of-flight momentum microscope, enabling simultaneous measurements across the full accessible energy and in-plane momentum ranges.

We first measure the equilibrium band structure and associated CDAD of unpumped graphene on a SiC substrate. The photoemission intensity along the direction perpendicular to Γ\Gamma-K through the Dirac point shows the linear dispersion of the π\pi bands (Figure 1d, left). Previous work has shown that static XUV CDAD is sensitive to the lattice pseudospin of graphene [38, 39, 40]. We thus measure energy- and momentum-resolved photoemission intensities while continuously modulating the XUV polarization state (by rotating a half-wave plate in front of a fixed quarter-wave plate) and extract the CDAD by Fourier-analyzing the polarization-modulated signal [10, 22] (see Figure S1).

Figure S1. Simultaneous extraction of CDAD and LDAD. (a) Photoemission intensity from graphene on SiC as a function of energy and momentum kyk_y along a cut through the Dirac point, measured without pump excitation. The dashed lines indicate the energy window used to extract the constant-energy contours (CECs) shown in (b,c). (b) CEC of the circular dichroism in the angular distribution (CDAD) around the K point, obtained from the imaginary parts of the Fourier coefficients of the periodically modulated photoemission signal recorded during continuous modulation of the XUV polarization state (see Methods). The CDAD exhibits antisymmetry with respect to ky=0k_y = 0. (c) CEC of the linear dichroism in the angular distribution (LDAD), obtained from the real parts of the Fourier coefficients of the same periodically modulated signal (see Methods). The LDAD exhibits antisymmetry about the K point along kxk_x. Both CECs display the symmetry patterns expected from the pseudospin texture of the Dirac cone, demonstrating that CDAD and LDAD can be reliably extracted simultaneously from a single measurement.

The measured dichroism (Figure 1d, right) changes sign between the two Dirac-cone branches and again upon crossing the Dirac energy. These features are consistent with the expected equilibrium pseudospin texture [39].

We now turn our attention to light-driven graphene. Figure 2a–Figure 2c shows the CDAD at zero pump–probe delay for a linearly polarized (s-polarized) pump and for σ+\sigma^{+}- and σ−\sigma^{-}-circularly polarized pump pulses (incident fluence of ∼3.5\sim3.5 mJ/cm2^{2}, i.e., ∼4000\sim4000 kV/cm). Upon s-polarized pump, time-reversal symmetry is preserved, leaving the dichroism essentially unchanged from its equilibrium behavior (Figure 1d), with no additional feature emerging around the Dirac point. In contrast, circularly polarized pumping induces an additional dichroic contribution near the Dirac point, whose sign reverses with the pump helicity (see red and blue arrows in Figure 2b–Figure 2c and Figure 2e–Figure 2f). As discussed above, the leading intrinsic contribution to the CDAD is proportional to the out-of-plane pseudospin component σz\sigma_z [34]. These new dichroic features therefore reflect the pump-induced modification of the pseudospin texture. Measurements at a lower pump fluence of 1.6 mJ/cm2^{2} (Figure S2) reveal no obvious helicity-dependent dichroic feature near the Dirac point, consistent with the reduced strength of the light-induced mass under weaker driving.

Figure 2. Pump-polarization-dependent XUV CDAD in light-driven graphene. (a) Measured CDAD at zero pump-probe delay for a linearly polarized (s-) pump, which retains the sign structure and symmetry of the static CDAD. (b)-(c) Measured CDAD for a right- (σ+\sigma^{+}) and left- (σ−\sigma^{-}) circularly polarized pump, respectively. For both pump helicities, the dichroism acquires a finite contribution close to the Dirac point that is absent for the s-polarized pump. Its sign changes with pump helicity: negative (blue arrow) for σ+\sigma^{+} and positive (red arrow) for σ−\sigma^{-}. The incident fluence is ∼3.5\sim3.5 mJ/cm2^{2}, i.e., ∼4000\sim4000 kV/cm. (d)-(f) Corresponding ab initio calculations of the CDAD for graphene driven by (d) s-polarized, (e) σ+\sigma^{+}, and (f) σ−\sigma^{-} pumps, with a drive amplitude of 4000 kV/cm. In contrast to Figure 1b, these simulations account for substrate effects and associated photoemission matrix elements.

Figure S2. Circular dichroism at low pump fluence. Circular dichroism in the angular distribution (CDAD) of graphene as a function of energy and momentum kyk_y along a cut through the Dirac point, measured at temporal overlap of the pump and probe pulses for a (a) right- (σ+\sigma^{+}) and a (b) left-circularly polarized (σ−\sigma^{-}) pump, at a reduced pump fluence of 1.6 mJ/cm21.6\ \mathrm{mJ}/\mathrm{cm}^{2}. No discernible pump-helicity dependence of the CDAD is observed near the Dirac point.

The precise energy and momentum dependence of these dichroic features is governed by the photoemission matrix elements. For graphene on SiC at a photon energy of 21.6 eV, matrix-element effects strongly redistribute the spectral weight between the lower and upper halves of the Dirac cone (Figure S3). As a result, in contrast to free-standing graphene (Figure 1b), the light-induced dichroic contributions do not exhibit a direct sign reversal across the Dirac energy. Instead, they appear as a single-sign feature: positive (red) for σ−\sigma^{-} pumping and negative (blue) for σ+\sigma^{+} pumping.

Figure S3. Polarization dependence of the photoemission intensity. (a-c) Measured photoemission intensity of graphene on SiC along kyk_y through the Dirac point at a photon energy of 21.6 eV, with (a) p-polarized XUV, (b) s-polarized XUV, and (c) integrated over all polarization states during continuous waveplate rotation (see Methods). In (a) and (b), the intensity is normalized to its maximum. In (c), the intensity at each QWP angle is normalized to its maximum in every energy row before integration, which reduces the matrix-element contribution. (d-f) Corresponding calculated photoemission intensity for p-polarized XUV (d), s-polarized XUV (e), and XUV integrated over all polarization states, including linear and circular polarizations (f). Each calculated map is normalized to its own maximum.

Our calculations capture this asymmetry (see Methods) and reproduce the measured CDAD for all three pump polarizations (Figure 2d–Figure 2f). The opposite signs of the dichroic response above and below the Dirac point will, however, become evident in the DD(Figure 3), as discussed below.

Figure 3. Spectroscopic signature of light-induced topological state. (a) Double-helicity dichroism (DD), obtained from the difference between the CDAD measured with σ+\sigma^{+} and σ−\sigma^{-} pump (Figure 2b,Figure 2c), which isolates the pump-helicity-odd component of the dichroism. (b) MDCs of the DD extracted above (E>EDE > E_D, red) and below (E<EDE < E_D, blue) the Dirac energy, together with Gaussian fits. (c) EDC of the DD extracted at the Dirac momentum kDk_D, together with a double-Gaussian fit. (d-f) Corresponding ab initio calculations of the DD (d), the MDCs above and below EDE_D (e), and the EDC at kDk_D (f), reproducing the localization of the DD around the Dirac point and its sign change across the Dirac energy.

To reveal the sign change of the dichroic response across the Dirac energy, we use the doubly differential observable DD (Figure 3), which removes the equilibrium dichroism and other contributions that are independent of the pump helicity, thereby isolating the helicity-odd response of the driven system. As shown in Figure 3a, the measured DD signal is positive above the Dirac energy and negative below it. This behavior is further highlighted by the normalized momentum distribution curves (MDCs) extracted from DD above and below the Dirac energy (Figure 3b), which exhibit a positive peak for E>EDE > E_D and a negative dip for E<EDE < E_D. The corresponding energy distribution curve (EDC), extracted at the Dirac momentum (Figure 3c), further confirms this sign reversal of DD across the Dirac energy.

The sign reversal of DD is consistent with the opposite signs of the light-induced out-of-plane pseudospin in the two bands. Our ab initio calculations (Figure 3d–Figure 3f), which account for both the Floquet dressing of the initial states and the photoemission matrix elements, reproduce the measured DD. Specifically, they capture the opposite signs of the MDCs above and below EDE_D (Figure 3e), and the sign reversal of the EDC at the Dirac momentum kDk_D (Figure 3f). This agreement supports the interpretation of the observed DD as a signature of the light-induced modification of the pseudospin texture of the Floquet topological insulating states.

Here, we go one step further by analyzing the link between the light-induced pseudospin texture, Berry curvature, and CDAD. The different momentum dependencies of the pseudospin and Berry curvature can be understood from the effective Floquet Hamiltonian Heff(k)=D(k)⋅σ^H_{\mathrm{eff}}(\mathbf{k}) = \mathbf{D}(\mathbf{k}) \cdot\hat{\boldsymbol{\sigma}}, where σ^\hat{\boldsymbol{\sigma}} is the vector of Pauli matrices acting on the A/B sublattice space and D(k)\mathbf{D}(\mathbf{k}) is the Dirac vector. The eigenenergies are ±∣D(k)∣\pm|\mathbf{D}(\mathbf{k})|, and the pseudospin expectation values of the two bands are aligned and anti-aligned with D(k)\mathbf{D}(\mathbf{k}), respectively. Near the Dirac points, the out-of-plane component of D\mathbf{D} is given by the light-induced mass ±mF\pm m_F, with opposite signs at the two valleys. The out-of-plane pseudospin is therefore proportional to mF/∣D(k)∣m_F/|\mathbf{D}(\mathbf{k})|, where ∣D(k)∣=(ℏvFk)2+mF2|\mathbf{D}(\mathbf{k})| = \sqrt{(\hbar v_F k)^2 + m_F^2}, vFv_F is the Fermi velocity, and k\mathbf{k} is measured relative to the respective KK or K′K' point. The Berry curvature of the two bands is given by Ω±(k)=∓12D^⋅(∂kxD^×∂kyD^)\Omega_\pm(\mathbf{k}) = \mp\frac{1}{2}\hat{\mathbf{D}} \cdot(\partial_{k_x}\hat{\mathbf{D}} \times\partial_{k_y}\hat{\mathbf{D}}), where D^=D/∣D∣\hat{\mathbf{D}} = \mathbf{D}/|\mathbf{D}|.

Figure 4a,Figure 4b compares the band-resolved out-of-plane pseudospin σz(k)\sigma_z(\mathbf{k}) and Berry curvature Ω(k)\Omega(\mathbf{k}), calculated from the Floquet Hamiltonian of the tight-binding model. Within the massive-Dirac description, their magnitudes vary as ∣σz(k)∣∝∣mF∣[(ℏvFk)2+mF2]−1/2|\sigma_z(\mathbf{k})| \propto|m_F|[(\hbar v_F k)^2 + m_F^2]^{-1/2} and ∣Ω(k)∣∝∣mF∣(ℏvF)2[(ℏvFk)2+mF2]−3/2|\Omega(\mathbf{k})| \propto|m_F|(\hbar v_F)^2[(\hbar v_F k)^2 + m_F^2]^{-3/2}. Away from the Dirac point, the Berry curvature thus decays as 1/k31/k^3, whereas the pseudospin decays only as 1/k1/k. This difference is illustrated in Figure 4c-Figure 4e. While the Berry curvature is strongly localized near the Dirac point, the pseudospin extends further along the bands. This is particularly useful for small light-induced gaps (Figure 4f), where scattering can obscure the gap without completely suppressing the dichroic signal.

Conclusions

Our results provide direct spectroscopic evidence for a Floquet topological insulating state in graphene, finally validating a landmark theoretical prediction that has long driven efforts to realize light-induced topological phases of matter. Rather than inferring topology indirectly from spectral signatures, our helicity-resolved photomission circular dichroic measurements directly access the quantum geometry of light-dressed wavefunctions. We resolve the pseudospin texture generated by the light-induced Haldane mass in excellent agreement with ab initio theory and simulations. These findings establish momentum-resolved wavefunction geometry as a direct probe of Floquet topology, robust even in the presence of scattering and heating. Furthermore, pump helicity controls the signs of the induced mass, Berry curvature, and orbital magnetization, enabling reversal of these properties on the optical-pulse timescale. This capability opens a route to writing, switching, and erasing topological states with femtosecond temporal resolution. Our results thus establish topology not merely as a static property of a quantum ground state, but as a dynamically controllable degree of freedom, opening new opportunities for ultrafast topological control in quantum materials.

Methods

Time- and polarization-resolved momentum microscopy The experiments were carried out using a polarization-controllable ultrafast extreme ultraviolet (XUV) beamline [42] connected to a momentum microscope end-station [10, 43, 44] at the Centre Lasers Intenses et Applications (CELIA). The laser source was a high-repetition-rate Yb fiber laser (166 kHz, 1030 nm, 135 fs, 50 W; Amplitude Laser Group). In the high-harmonic generation (HHG) arm, a substantial fraction of the fundamental laser output was frequency-doubled in a BBO crystal, yielding several watts of 515 nm radiation. The resulting beam was resized and spatially shaped into an annular profile before being focused into an argon gas jet, to generate XUV radiation through HHG. After the gas jet, residual 515 nm radiation was suppressed through successive spatial filtering stages employing pinholes. The ninth harmonic of the 515 nm driving field, corresponding to a photon energy of 21.6 eV, was selected using reflections from Sc/SiC multilayer XUV mirrors (NTTAT) in combination with a 200 nm-thick Sn filter (Luxel).

The infrared (IR) pump pulses were obtained by extracting a small fraction of the fundamental laser output (1030 nm, 135 fs). The polarization was adjusted using a motorized quarter-wave plate positioned immediately upstream of the recombination chamber. The pump and probe beams were combined collinearly using a drilled mirror and focused onto the sample. The overall temporal resolution was approximately 150 fs. The IR pump fluence was determined in situ using the photoemission electron microscopy (PEEM) mode of our time-of-flight momentum microscope, which enables real-space imaging of photoelectrons emitted from the sample surface. Because photoemission induced by the 1.2 eV IR pulses is a multiphoton process, the measured PEEM image does not directly reproduce the spatial profile of the IR beam. Instead, it reflects the beam profile modified by the effective nonlinearity of the photoemission process. To account for this effect, we determined the effective nonlinearity by measuring the total photoemission intensity as a function of the IR pump power. Using the extracted nonlinearity, we reconstructed the actual IR beam profile from the PEEM images, enabling an accurate in situ determination of the pump fluence.

The graphene sample (Graphene-SiC) was mounted on a motorized six-axis hexapod, which enabled precise sample positioning and alignment in the main chamber. Measurements were conducted at room temperature under a base pressure of 2×10−102 \times10^{-10} mbar. Time-resolved photoemission data were acquired with a custom-built time-of-flight momentum microscope featuring an advanced front lens with multiple operating modes (GST mbH) [43, 44]. For XUV circular dichroism in the angular distribution (CDAD) measurements, an all-reflective XUV quarter-wave plate was placed in the probe beam path. During acquisition, a visible-light half-wave plate (515 nm), positioned upstream of the HHG chamber (before the XUV quarter-wave plate), was continuously rotated while the photoemission signal was recorded. Each detected electron event was correlated with the instantaneous angular position of the half-wave plate. Owing to the reflection of the XUV beam from the quarter-wave plate, one complete 360∘360^\circ rotation of the driving-beam half-wave plate generated four successive cycles of p-, left-circularly polarized (LCP), s-, right-circularly polarized (RCP), and p-polarized XUV radiation. To determine the dichroic response, we performed a Fourier analysis independently for each energy-momentum voxel, resolving the photoemission intensity as a function of the XUV polarization angle. The real and imaginary components of the resulting oscillatory signal were used to quantify the linear and circular dichroism in XUV photoemission, respectively. Raw single-event data were processed using an open-source data-analysis workflow [45, 46], which converts the event-level measurements into calibrated, binned hypervolumes with the required dimensions. Further details of the experimental apparatus and the CDAD extraction using Fourier analysis are provided in Ref. 10.

Graphene sample The 7 mm×7 mm7\ \mathrm{mm} \times7\ \mathrm{mm} monolayer epitaxial graphene sample on SiC was purchased from Graphene-SiC. The graphene was synthesized by graphitization of the Si-face of 4H–SiC [47]. The sample is predominantly covered by monolayer graphene, with additional bilayer graphene present in the form of stripe-like domains. According to the manufacturer, optical inspection confirmed a monolayer coverage exceeding 95%, with bilayer graphene inclusions at 5%. The sample was annealed at 400°C for 2 hours to prepare a clean surface before photoemission measurements.

Theoretical calculations The simulations of the trARPES signal at pump-probe overlap were carried out using the non-equilibrium Green’s function formalism, specifically the Floquet steady-state approach described in ref. 48. In brief, the photoemission intensity is computed from

I(k,E)=Im⁡∑nn′Mn(k,E)Gnn′<(k,ωprobe−E)Mn′∗(k,E)(1)I(\mathbf{k}, E) = \operatorname{Im} \sum_{nn'} M_n(\mathbf{k}, E)G^{<}_{nn'}(\mathbf{k}, \omega_{\mathrm{probe}} - E)M_{n'}^{*}(\mathbf{k}, E) \tag*{(1)}

where Mn(k,E)M_n(\mathbf{k}, E) are the photoemission matrix elements, and where Gnn′<(k,ω)G^{<}_{nn'}(\mathbf{k}, \omega) is the period-averaged lesser Green’s function. The kinetic energy of the photoelectrons is denoted by EE, while ωprobe\omega_{\mathrm{probe}} is the probe photon energy. To calculate the Green’s function, two ingredients are required: (i) the light-matter coupling of the pump pulse, and (ii) a dissipative bath defining the steady state and the effective occupations. The latter is described by a wide-band fermionic bath with a coupling strength of γ=0.25\gamma= 0.25 eV for the conduction band and γ=0.125\gamma= 0.125 eV for the valence band. A chemical potential of μ=0.5\mu= 0.5 eV is used to emulate the doping of the sample. For (i), we computed the velocity matrix elements vnn′(k)=i[H^KS(k),r^]nn′v_{nn'}(\mathbf{k}) = i[\hat{H}_{\mathrm{KS}}(\mathbf{k}), \hat{\mathbf{r}}]_{nn'} directly from the Kohn-Sham (KS) Hamiltonian after performing density-functional theory (DFT) calculations of the band structure. In the velocity gauge, this procedure yields the time-dependent Hamiltonian

Hnn′(k,t)=(εn(k)+q2A(t)22)δnn′−qA(t)⋅vnn′(k)(2)H_{nn'}(\mathbf{k},t) = \left(\varepsilon_n(\mathbf{k}) + \frac{q^2 A(t)^2}{2}\right)\delta_{nn'} - q\mathbf{A}(t)\cdot\mathbf{v}_{nn'}(\mathbf{k}) \tag*{(2)}

where A(t)=(E0/ωpump)Re⁡[iu^e−iωpumpt]\mathbf{A}(t) = (E_0/\omega_{\mathrm{pump}})\operatorname{Re}[i\hat{\mathbf{u}}e^{-i\omega_{\mathrm{pump}}t}] is the pump vector potential. The effective polarization vector u^\hat{\mathbf{u}} was computed as the superposition of the incoming and reflected fields at the interface of the SiC substrate under the experimental angle of incidence, using a dielectric constant of ϵSiC(ωpump)=6.7\epsilon_{\mathrm{SiC}}(\omega_{\mathrm{pump}})=6.7. The time-dependent Hamiltonian (2) is then used to construct the Floquet Hamiltonian and Gnn′(k,ω)G_{nn'}(\mathbf{k},\omega) (for details see ref. 48). We also include laser-assisted photoemission (LAPE) in the Green’s function.

The photoemission matrix elements Mn(k,E)M_n(\mathbf{k},E) for free-standing graphene are computed from first principles using the approach described in ref. 49. Compared to the experiments, we notice some deviations in the equilibrium ARPES signal, which are induced by the SiC substrate. In particular, the photoemission intensity is strongly reduced for p-polarization; in the experimental geometry, this is only possible for an almost vanishing dielectric constant ϵSiC(ωprobe)\epsilon_{\mathrm{SiC}}(\omega_{\mathrm{probe}}). The most likely explanation is a bulk plasmon resonance, which has been observed in electron energy-loss spectroscopy experiments [50] at ∼22\sim22 eV. A simple density counting argument also yields ωp=4πn≃22\omega_p=\sqrt{4\pi n}\simeq22 eV. We fix ϵSiC(ωprobe)=0.02+0.1i\epsilon_{\mathrm{SiC}}(\omega_{\mathrm{probe}})=0.02+0.1i and compute the probe polarization from the Fresnel equations (superposition of incoming and reflected fields). Additionally, we scale the photoemission matrix element for p-polarization, Mcp(k,E)→rMcp(k,E)M_c^p(\mathbf{k},E)\to rM_c^p(\mathbf{k},E), r=12r=\frac{1}{\sqrt{2}}. To account for the reduced intensity of circular dichroism results above the Dirac point, we further scale all intensities with a smooth step function going from 1 below EDE_D to 1/2 above EDE_D. This model for matrix elements yields polarization-resolved equilibrium spectra in good agreement with the experiments (see Figure S3), including the circular dichroism.

For the results presented in Figure 4, we used the two-band Dirac model HD(k)=D(k)⋅σ^H_D(\mathbf{k})=\mathbf{D}(\mathbf{k})\cdot\hat{\boldsymbol{\sigma}}, where D(k)=(−τℏvFkx,ℏvFky,0)\mathbf{D}(\mathbf{k})=(-\tau\hbar v_F k_x,\hbar v_F k_y,0). Here, τ\tau is the valley index, while vFv_F denotes the Fermi velocity. We then constructed the time-dependent Hamiltonian HD(k,t)=HD(k−qA(t))H_D(\mathbf{k},t)=H_D(\mathbf{k}-q\mathbf{A}(t)) via the Peierls substitution and constructed the corresponding Floquet Hamiltonian Ham,a′m′(k)\mathcal{H}_{am,a'm'}(\mathbf{k}), where aa, a′a' (mm, m′m') denote the sublattice sites (Floquet indices). Diagonalizing the Floquet Hamiltonian yields the quasienergies ελ(k)\varepsilon_{\lambda}(\mathbf{k}) with corresponding eigenvectors Wam,λ(k)W_{am,\lambda}(\mathbf{k}), from which we compute the pseudospin σz(k)=∣WAm,λ(k)∣2−∣WBm,λ(k)∣2\sigma_z(\mathbf{k}) = \lvert W_{Am,\lambda}(\mathbf{k})\rvert^2 - \lvert W_{Bm,\lambda}(\mathbf{k})\rvert^2 for band and Floquet indices λ\lambda, mm corresponding to the effective valence band. Similarly, the Berry curvature is computed via the sum-over-bands definition Curvature

Ωλ(k)=2Im⁡(∑λ′≠λvλλ′,x(k)vλ′λ,y(k)−vλλ′,y(k)vλ′λ,x(k)(ελ′(k)−ελ(k))2)(3)\Omega_{\lambda}(\mathbf{k}) = 2\operatorname{Im}\left(\sum_{\lambda'\ne\lambda}\frac{v_{\lambda\lambda',x}(\mathbf{k})v_{\lambda'\lambda,y}(\mathbf{k}) - v_{\lambda\lambda',y}(\mathbf{k})v_{\lambda'\lambda,x}(\mathbf{k})}{\left(\varepsilon_{\lambda'}(\mathbf{k})-\varepsilon_{\lambda}(\mathbf{k})\right)^2}\right) \tag*{(3)}

where vλλ′,μ(k)=∑m,aa′Wam,λ∗(k)vaa′,μ(k)Wa′m,λ(k)v_{\lambda\lambda',\mu}(\mathbf{k}) = \sum_{m,aa'} W^*_{am,\lambda}(\mathbf{k})v_{aa',\mu}(\mathbf{k})W_{a'm,\lambda}(\mathbf{k}). The velocity matrix elements vaa′,μ(k)v_{aa',\mu}(\mathbf{k}) are calculated from the Dirac model here.

Data Availability

The data that support the findings of this article will be openly available on Zenodo before publication .

Acknowledgement

We thank Charlotte Sanders for insightful discussions at the project’s initial stage, and Michael A. Sentef for fruitful discussions during the final phase. We thank Nikita Fedorov, Romain Delos, Pierre Hericourt, Rodrigue Bouillaud, Laurent Merzeau, and Frank Blais for technical assistance. We thank Baptiste Fabre for implementing and maintaining the data binning code. We acknowledge the financial support of the IdEx University of Bordeaux/Grand Research Program “GPR LIGHT”. This work is part of the ULTRAFAST and TORNADO projects of PEPR LUMA and was supported by the French National Research Agency, as a part of the France 2030 program, under grants ANR-23-EXLU-0002 and ANR-23-EXLU-0004. We acknowledge support from ERC Starting Grant ERC-2022-STG No.101076639, Quantum Matter Bordeaux, AAP CNRS Tremplin, and AAP SMR from Université de Bordeaux. S.M. and W.B. acknowledge support from the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation), project number 567214325, and by the Lower Saxony Ministry of Science and Culture (MWK) and the VolkswagenStiftung through the zukunft.niedersachsen program ”Wissenschaftsräume” (Wissenschaftsraum ElLiKo, ”Elektronen–Licht–Kontrolle”). W.B. acknowledges support from the Dorothea Schlözer Postdoctoral Programme for Women. S.F. acknowledges funding from the European Union’s Horizon Europe research and innovation programme under the Marie Skłodowska-Curie 2024 Postdoctoral Fellowship No 101198277 (TopQMat). U.D. acknowledges funding from the European Union’s Horizon Europe research and innovation programme under the Marie Skłodowska-Curie grant No HORIZON-MSCA-2023-DN-01 101169225 - SPARKLE. A.C.G. and M.S. acknowledge funding from the Swiss National Science Foundation (SNSF), grant no. 10003275. P.H. acknowledges support by Novo Nordisk Foundation (Grant number NNF23OC0085585). Funded by the European Union. Views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union. Neither the European Union nor the granting authority can be held responsible for them.

Author contributions

S.B. and M.S. conceived the idea. S.F. and S.B. planned and performed the experiments. U.D., W.B. and Y.M. assisted with the experiments. S.F., Y.M. and S.B. analyzed the experimental data. D.D. and S.P. helped maintain the laser system. M.S. and A.C.G. developed the theory and associated interpretations. S.F., S.B., A.C.G. and M.S. provided the initial interpretation of the results. S.F., A.C.G., W.B., P.H., M.R., S.M. and S.B. discussed the results, analysis, and interpretation. S.F., S.B., A.C.G. and M.S. wrote the first draft of the manuscript with input from U.D., D.D., W.B., P.H., M.R., S.M., and Y.M.

Competing Interests

The authors declare that they have no competing financial interests.

Supplementary Information: Observation of light-induced topological state in graphene

Simultaneous extraction of CDAD and LDAD. The static CDAD and LDAD of graphene shown in Figure S1 are extracted from a single measurement in which the XUV polarization state is continuously modulated (see Methods). This single-scan approach avoids systematic errors that can arise when CDAD and LDAD are measured sequentially under nominally identical conditions using separately acquired datasets for different polarization states. It underpins all static and time-resolved dichroism measurements presented in the main text. The corresponding constant-energy contours (Figure S1,Figure S1) exhibit the antisymmetry expected from the pseudospin texture of the Dirac cone [41], confirming the reliable and simultaneous extraction of both observables from the same dataset.

Fluence dependence of the pump-helicity-dependent dichroism. We investigate the CDAD at a reduced pump fluence of 1.6 mJ/cm21.6\ \mathrm{mJ}/\mathrm{cm}^{2}, compared with the 3.5 mJ/cm23.5\ \mathrm{mJ}/\mathrm{cm}^{2} used in the main text (Figure S2).

At this lower fluence, no discernible difference is observed between the CDAD measured using σ+\sigma^{+} and σ−\sigma^{-} pump pulses near the Dirac point, in contrast to the pronounced helicity dependence observed in Figure 2 of the main text. This behavior is consistent with the reduced magnitude of the light-induced mass term at weaker driving strengths. Since this term scales quadratically with the pump field amplitude, its effect is expected to fall below our detection sensitivity at sufficiently low pump fluences.

XUV polarization dependence of the photoemission intensities Figure S3 shows the photoemission intensity of graphene on SiC along kyk_y through the Dirac point, recorded at a photon energy of 21.6 eV for different XUV probe polarizations. The distribution of spectral weight between the upper and lower Dirac cones depends strongly on the probe polarization. For p-polarized XUV (Figure S3), the intensity above the Dirac point is strongly suppressed, with most of the spectral weight concentrated in the lower cone. For s-polarized XUV (Figure S3), the intensity is distributed more evenly between the two cones, and both branches remain visible across the Dirac point. Since circularly polarized XUV contains a p-polarized component, a similar imbalance is expected in the CDAD measurements. This is consistent with the predominantly single-sign CDAD observed near the Dirac point (Figure 2), in contrast to the calculations for free-standing graphene (Figure 1b). The observed behavior is reproduced by the calculations when the modified photoemission matrix elements are taken into account.

The suppression of the upper cone also leads to an intensity depletion near the Dirac point that resembles a gap. This feature is substantially more pronounced for p-polarized than for s-polarized XUV, indicating a strong influence of photoemission matrix elements. To mitigate these effects, we integrate the photoemission intensity over the range of polarization states sampled during continuous rotation of the half-wave plate placed before the XUV quarter-wave plate (see Methods). Before integration, the intensity recorded at each QWP angle is normalized to its maximum within each energy row. This procedure reduces the polarization- and energy-dependent intensity modulation introduced by the photoemission matrix elements while preserving the band dispersion. In the resulting spectrum (Figure S3), the intensity distribution becomes more continuous across the Dirac point, and the gap-like depletion is further suppressed. We therefore attribute this feature primarily to photoemission matrix-element effects, although a small gap below our experimental resolution cannot be ruled out.

Our simulations of the equilibrium ARPES spectrum, using the photoemission matrix elements Mn(k,E)M_n(\mathbf{k}, E) for free-standing graphene and applying the scaling explained in the Methods section, are in good qualitative agreement with the experimental spectra (Figure S3–Figure S3). In particular, the suppressing of the photoemission intensity above EDE_D for p polarization is reproduced well. We use the thus calibrated photoemission matrix elements consistently for all calculated results in the main text.

Effect of a trivial gap on pseudospin, Berry Curvature, and dichroisms. In the main text, the pump-helicity-dependent CDAD near the Dirac point is interpreted as a signature of the light-induced pseudospin texture. Since graphene on SiC can also exhibit a trivial gap, we examine whether our main interpretation, namely, that the DD probes the pseudospin texture associated with a Floquet topological insulator, remains robust in the presence of such a trivial gap already present at equilibrium.

For the ab initio Hamiltonian, a gap can be introduced by modifying the Hamiltonian

H~nn′(k)=εn(k)δnn′+Δ2i∣C(k)∣[vx(k),vy(k)]nn′(4)\widetilde{H}_{nn'}(\mathbf{k}) = \varepsilon_n(\mathbf{k})\delta_{nn'} + \frac{\Delta}{2i\lvert C(\mathbf{k})\rvert}[v_x(\mathbf{k}),v_y(\mathbf{k})]_{nn'} \tag*{(4)}

where ∣C(k)∣\lvert C(\mathbf{k})\rvert is the largest eigenvalue of the commutator [vx(k),vy(k)][v_x(\mathbf{k}),v_y(\mathbf{k})]. For the TB model, the gap is introduced by replacing the diagonal elements by Δ\Delta (−Δ-\Delta) for the A (B) sublattice site. In both models, the system exhibits a gap Egap=∣Δ∣E_{\mathrm{gap}}=\lvert\Delta\rvert and a trivial mass of ∣Δ∣/2\lvert\Delta\rvert/2. The sign of Δ\Delta is determined by the spatial arrangement of the graphene sheet on the SiC substrate. Fixing a value for Δ\Delta, a net CDAD is observed even in equilibrium, which is not observed in the experiments. Thus, we can assume that domains with different orientation are present, with roughly equal contributions to the total photoemission signal. Mathematically, this scenario is modeled by performing calculations for ±Δ\pm\Delta and averaging the results. Eq. (4) also allows us to study pure Floquet effects within the high-frequency perturbation theory by identifying Δ\Delta with the Floquet mass mFm_{\mathrm{F}}.

Fig. S4 shows the light-induced pseudospin and Berry curvature in the presence of the trivial gap. For small field strength, the pseudospin is pinned to zero at ky=0k_y = 0 relative to the Dirac point; away for ky=0k_y = 0, the behavior closely resembles the case of Egap=0E_{\mathrm{gap}} = 0. As soon as the light-induced gap exceeds the trivial one, pseudospin approaches 1, and the profile of σz(k)\sigma_z(\mathbf{k}) becomes very similar to the pristine graphene case (Figure S4,Figure S4). The Berry curvature exhibits similar behavior (Figure S4,Figure S4).

Figure S4. Impact of trivial substrate-induced gap. Light-induced pseudospin σz(k)\sigma_z(\mathbf{k}) and Berry curvature for the gapped model, for Egap=30E_{\mathrm{gap}} = 30 meV (a,b) and Egap=60E_{\mathrm{gap}} = 60 meV (c,d). The result were obtained from the high-frequency expansion of the TB model.

Next, we investigated the impact of the substrate-induced gap onto the photoemission signal and the circular dichroism. Figure S5 shows the calculated CDAD maps and corresponding DD for Egap=60E_{\mathrm{gap}} = 60 meV (a–c) and 240 meV (d–f), both for the peak field strength of E0=4000E_0 = 4000 kV/cm used for all simulations in the main text. For the smaller gap of Egap=60E_{\mathrm{gap}} = 60 meV, the CDAD and the DD are very similar to the pristine case, consistent with behavior the light-induced pseudospin and Berry curvature (Figure S4,Figure S4). These results show that the dichroic features associated with the pseudospin and Berry curvature do not require the trivial equilibrium gap to vanish. Increasing the gap beyond the Floquet gap, the dichroic signal near the Dirac point is visibly suppressed in the CDAD maps (Figure S5,Figure S5), while the DD (Figure S5) develops an extended region with a vanishing signal. Neither behavior is observed in our experimental data, suggesting that our samples do not lie in the regime of such a large trivial gap.

Figure S5. Effect of trivial band gap on CDAD and DD. Calculated XUV CDAD along kyk_y through the Dirac point for pump helicity σ+\sigma^{+} (a,d) and σ−\sigma^{-} (b,e), and the corresponding double difference (DD) between the two pump helicities (c,f). Panels a–c correspond to a trivial gap Egap=60E_{\mathrm{gap}} = 60 meV and panels d–f to Egap=240E_{\mathrm{gap}} = 240 meV, for a drive field of 4000 kV/cm. The dichroic features near the Dirac point, associated with the pseudospin and Berry-curvature texture, are present for both gap values, but are fading and accompanied by a large region of vanishing signal in the DD for the larger gap.

Topological origin of the dichroism Because our measurement geometry combines a large angle of incidence with a high-photon-energy XUV probe, we performed extensive sanity checks to confirm that the observed dichroism genuinely originates from light-induced pseudospin texture rather than from a trivial geometric or matrix-element artifact of the oblique, elliptically-polarized drive. First, replacing the physical pump with its closest linear-polarization proxy of identical in-plane field magnitude removes all helicity content, and the resulting proxy dichroism is practically zero (Figure S6). This confirms that the dichroism is driven by circularity of the pump, not by the incidence geometry itself. For a true circular pump, the details of the geometry and the dielectric constant merely determine the overall scale but not the signature of the topological state in the DD, as demonstrated by the exact same features observed for various conditions in Figure S6–Figure S6).

Figure S6. Topological origin of the DD. a Simulated DD, using the methodology described in the Methods section, where the effective pump polarization in the (x,y)(x,y) plane was replaced by its closest linear proxy. b–d: Simulated DD for circular pump polarization (with respect to the vacuum propagation direction) for different dielectric constants ϵ\epsilon and angle of incidence θ\theta. d corresponds to the conditions in the experiments. e Quantitative evaluation of the strength of the DD, comparing the different geometry conditions from b–d and the linear proxy case from a. f: Maximum strength of the DD from the full simulation as a function of the peak field strength E0E_0.

Second, the extracted DD follows the field scaling expected for a second-order (helicity-odd) light–matter process, scaling quadratically with pump field at weak drive (Figure S6).

Finally, we isolate the purely helicity-odd, second-order Floquet term responsible for the light-induced Berry curvature – while discarding all other contributions to the light–matter coupling – by high-frequency Floquet perturbation theory, Eq. (4) with Δ=mF\Delta= m_{\mathrm{F}}. Computing the corresponding effective Green’s function

G~R(ω)=[ω−H~(k)+iη]−1,G~<(ω)=if(ω)[G~R(ω)−G~R(ω)†],(5)\widetilde{G}^{\mathrm{R}}(\omega) = [\omega- \widetilde{H}(\mathbf{k}) + i\eta]^{-1}, \quad\widetilde{G}^{<}(\omega) = if(\omega)[\widetilde{G}^{\mathrm{R}}(\omega) - \widetilde{G}^{\mathrm{R}}(\omega)^\dagger], \tag*{(5)}

where f(ω)f(\omega) is the Fermi-Dirac distribution at equilibrium, and inserting into Eq. (1) in the main text yields the photoemission spectrum for the effective Floquet Hamiltonian. Figure S7 demonstrates that the DD from the full simulation can be reproduced almost quantitatively by the pure Floquet model, with a correlation of C(full, pure Floquet) = 0.927 in the full considered energy-momentum region, and C(full, pure Floquet) = 0.985 for the region close to the Dirac point (black dashed box in Figure S7).

Figure S7. Pure Floquet origin of the DD. a DD for the full simulation of the photoemission spectrum from the main text. b DD for the effective Floquet model (S2) derived from the high-frequency Floquet perturbation theory. We computed the correlation between the full simulation and the effective model for (i) the full energy-momentum region, and (ii) for the region in the vicinity of the Dirac point, indicated by the black dashed box.

Taken together, these checks rule out incidental geometric contributions and establish that the observed helicity-reversal dichroism is the direct spectroscopic fingerprint of the light-induced Berry curvature at the Dirac point.

References

  1. [1]Oka, T. & Aoki, H. Photovoltaic Hall effect in graphene. Phys. Rev. B 79, 081406 (2009).DOI
  2. [2]Oka, T. & Kitamura, S. Floquet Engineering of Quantum Materials. Annual Review of Condensed Matter Physics 10, 387–408 (2019).DOI
  3. [4]Haldane, F. D. M. Model for a Quantum Hall Effect without Landau Levels: Condensed-Matter Realization of the ”Parity Anomaly”. Phys. Rev. Lett. 61, 2015–2018 (1988).pubmed.ncbi.nlm.nih.gov/10038961
  4. [5]McIver, J. W. et al. Light-induced anomalous Hall effect in graphene. Nature Physics 16, 38–41 (2020).
  5. [6]Lesko, D. M. B. et al. Optical control of electrons in a Floquet topological insulator. Nat. Phys. 22, 1607–1614 (2026).DOI
  6. [7]Merboldt, M. et al. Observation of Floquet states in graphene. Nat. Phys. 21, 1093–1099 (2025).
  7. [8]Choi, D. et al. Observation of Floquet–Bloch states in monolayer graphene. Nat. Phys. 21, 1100–1105 (2025).DOI
  8. [9]Wang, F. et al. Observation of Floquet-induced gap in graphene. Nat. Mater. 25, 1140–1145 (2026).DOI
  9. [10]Fragkos, S. et al. Time- and polarization-resolved extreme ultraviolet momentum microscopy. Review of Scientific Instruments 96, 115201 (2025).arxiv.org/abs/2507.02371
  10. [11]de la Torre, A. et al. Colloquium: Nonthermal pathways to ultrafast control in quantum materials. Rev. Mod. Phys. 93, 041002 (2021).DOI
  11. [12]Basov, D. N., Averitt, R. D. & Hsieh, D. Towards properties on demand in quantum materials. Nature Materials 16, 1077–1088 (2017).DOI
  12. [13]Jotzu, G. et al. Experimental realization of the topological Haldane model with ultracold fermions. Nature 515, 237–240 (2014).DOI
  13. [14]Rechtsman, M. C. et al. Photonic Floquet topological insulators. Nature 496, 196–200 (2013).
  14. [15]Wang, Y. H., Steinberg, H., Jarillo-Herrero, P. & Gedik, N. Observation of Floquet-Bloch States on the Surface of a Topological Insulator. Science 342, 453–457 (2013).
  15. [16]Mahmood, F. et al. Selective scattering between Floquet–Bloch and Volkov states in a topological insulator. Nature Physics 12, 306–310 (2016).DOI
  16. [17]Zhou, S. et al. Pseudospin-selective Floquet band engineering in black phosphorus. Nature 614, 75–80 (2023).DOI
  17. [18]Zhou, S. et al. Floquet Engineering of Black Phosphorus upon Below-Gap Pumping. Phys. Rev. Lett. 131, 116401 (2023).DOI
  18. [19]Bao, C. et al. Manipulating the symmetry of photon-dressed electronic states. Nature Communications 15, 10535 (2024).DOI
  19. [20]Ito, S. et al. Build-up and dephasing of Floquet–Bloch bands on subcycle timescales. Nature 616, 696–701 (2023).DOI
  20. [21]Bao, C. et al. Light-induced ultrafast glide-mirror symmetry breaking in black phosphorus. ACS nano 18, 32038–32044 (2024).DOI
  21. [22]Fragkos, S. et al. Floquet-Bloch valleytronics. Nat. Commun. 16, 5799 (2025).DOI
  22. [23]Chassot, F. et al. Floquet topological state induced by light-driven band inversion in SnTe. Nat. Phys. 22, 1282–1286 (2026).DOI
  23. [24]Aeschlimann, S. et al. Survival of Floquet–Bloch States in the Presence of Scattering. Nano Letters 21, 5028–5035 (2021).
  24. [25]Breusing, M. et al. Ultrafast nonequilibrium carrier dynamics in a single graphene layer. Phys. Rev. B 83, 153410 (2011).DOI
  25. [26]Brida, D. et al. Ultrafast collinear scattering and carrier multiplication in graphene. Nature Communications 4, 1987 (2013).DOI
  26. [27]Heide, C. et al. Electronic coherence and coherent dephasing in the optical control of electrons in graphene. Nano Lett. 21, 9403–9409 (2021).DOI
  27. [28]Sato, S. A. et al. Floquet states in dissipative open quantum systems. J. Phys. B At. Mol. Opt. Phys. 53, 225601 (2020).
  28. [29]Broers, L. & Mathey, L. Detecting light-induced Floquet band gaps of graphene via trARPES. Phys. Rev. Res. 4, 013057 (2022).DOI
  29. [30]Sato, S. A. et al. Microscopic theory for the light-induced anomalous Hall effect in graphene. Phys. Rev. B 99, 214302 (2019).DOI
  30. [31]Balog, R. et al. Bandgap opening in graphene induced by patterned hydrogen adsorption. Nature Materials 9, 315–319 (2010).DOI
  31. [32]Hunt, B. et al. Massive Dirac Fermions and Hofstadter Butterfly in a van der Waals Heterostructure. Science 340, 1427–1430 (2013).DOI
  32. [33]Gorbachev, R. V. et al. Detecting topological currents in graphene superlattices. Science 346, 448–451 (2014).
  33. [34]Schüler, M. et al. How circular dichroism in time- and angle-resolved photoemission can be used to spectroscopically detect transient topological states in graphene. Phys. Rev. X 10, 041013 (2020).
  34. [35]Kim, Y. et al. Visualizing Berry curvature in a Floquet-Chern insulator (2026). URL https://arxiv.org/abs/2609.17500.
  35. [36]Mikami, T. et al. Brillouin-Wigner theory for high-frequency expansion in periodically driven systems: Application to Floquet topological insulators. Phys. Rev. B 93, 144307 (2016).DOI
  36. [37]Sentef, M. A. et al. Theory of Floquet band formation and local pseudospin textures in pump-probe photoemission of graphene. Nature Communications 6, 7047 (2015).DOI
  37. [38]Liu, Y., Bian, G., Miller, T. & Chiang, T.-C. Visualizing Electronic Chirality and Berry Phases in Graphene Systems Using Photoemission with Circularly Polarized Light. Phys. Rev. Lett. 107, 166803 (2011).
  38. [39]Gierz, I., Lindroos, M., Höchst, H., Ast, C. R. & Kern, K. Graphene Sublattice Symmetry and Isospin Determined by Circular Dichroism in Angle-Resolved Photoemission Spectroscopy. Nano Letters 12, 3900–3904 (2012).DOI
  39. [40]Hwang, C. et al. Direct measurement of quantum phases in graphene via photoemission spectroscopy. Physical Review B 84, 125422 (2011).DOI
  40. [41]Schüler, M. et al. Local Berry curvature signatures in dichroic angle-resolved photoelectron spectroscopy from two-dimensional materials. Science Advances 6, eaay2730 (2020).DOI
  41. [42]Comby, A. et al. Ultrafast polarization-tunable monochromatic extreme ultraviolet source at high-repetition-rate. Journal of Optics 24, 084003 (2022).DOI
  42. [43]Tkach, O. & Schönhense, G. Multimode objective lens for momentum microscopy and XPEEM: Theory. Ultramicroscopy 276, 114167 (2025).DOI
  43. [44]Tkach, O. et al. Multimode objective lens for momentum microscopy and x-ray photoemission electron microscopy: Experiments. Rev. Sci. Instrum. 97, 033703 (2026).
  44. [45]Xian, R. P. et al. An open-source, end-to-end workflow for multidimensional photoemission spectroscopy. Scientific Data 7, 442 (2020).
  45. [46]Xian, R. P., Rettig, L. & Ernstorfer, R. Symmetry-guided nonrigid registration: The case for distortion correction in multidimensional photoemission spectroscopy. Ultramicroscopy 202, 133 – 139 (2019).DOI
  46. [47]Emtsev, K. V. et al. Towards wafer-size graphene layers by atmospheric pressure graphitization of silicon carbide. Nature Materials 8, 203–207 (2009).DOI
  47. [50]Yang, S., Nakagawa, Y., Kondo, M. & Shibayama, T. Electron energy-loss spectroscopic evaluation of depth-dependent swelling of He+ ion-irradiated 4H-SiC correlated with defect type. Journal of Applied Physics 127, 175106 (2020).DOI

Paper details

vlnc:W_tqpfaxtk
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents