WorksPhysics and Astronomy
Controlling photo-excited electron-spin by light-polarization in ultrafast-pumped altermagnets
Abstract
Altermagnets (AMs) constitute a novel class of spin-compensated materials in which the symmetry connecting opposite-spin sublattices involves a spatial rotation. Here, we uncover a set of unique non-linear, light-driven properties that set AMs apart from traditional ferro- and antiferromagnets. We demonstrate theoretically that the polarization of an electromagnetic pulse that photo-excites electrons and holes in an AM, controls the spin orientation of these non-equilibrium charge carriers. For a d-wave AM model and a prototype material, we show that very large post-pump spin polarizations may be attained by exploiting resonances. We show that this protocol also allows, in an AM, to directly probe the spin splitting of the electronic states in energy and momentum space. Thus, it can be used to identify and characterize altermagnetic materials via ultrafast pump-probe Kerr/Faraday spectroscopy or spin- and time-resolved ARPES. This opens up the possibility of devising ultrafast optical switches of non-equilibrium spin-polarization, finely tunable by adjusting the pump-pulse characteristics.
Altermagnetism has recently emerged as a new type of magnetic ordering, distinct from ferro- and antiferromagnetism. Similarly to antiferromagnets (AFMs), the net magnetization of altermagnets (AMs) vanishes by symmetry. They differ from AFMs because the enforcing symmetry that connects the two magnetic sub-lattices is not merely an inversion or translation, but also involves a rotation [1, 2]. This leads to a breaking of the spin degeneracy of the electronic states in their non-relativistic band structure with an energy scale set by the local exchange field, which is generally much larger than the relativistic spin-orbit-coupling energy scale. A large number of AMs have already been identified [3, 4, 5, 6, 7], and this number continues to grow. Various linear response properties of AMs have been recognized that may render them of practical interest, for instance, as spin current generators relevant for spintronics [1, 2]. Also, the recently developed Landau theory of altermagnetism [8] allows to relate the formation of altermagnetic order directly to other key linear response properties such as anomalous Hall conductivity [1, 9], Edelstein response [10, 11] and piezomagnetic [12, 13, 14], magneto-optic, and magneto-elastic [15] effects. In particular, the linear magneto-optical response in the presence of an external magnetic field, which involves momentum- and spin-dependent matrix elements, is sensitive to the presence of altermagnetism and depends on the orientation of the ground state Néel vector [16].
Here, we take a step beyond the linear responses of AMs and consider their non-linear, light-driven properties. In general, the possibility of controlling the physical properties of solids with ultrafast electromagnetic (EM) pulses is a fascinating goal at the core of a broad research field [17, 18, 19, 20, 21, 22, 23]. In this Letter, we set out to determine how the features of a strong EM pulse (its polarization, frequency, amplitude, and duration) affect the post-pump photo-induced charge carriers (their density, spin and momentum) in AMs. Using the Dynamical Projective Operator Approach (DPOA) [23, 24, 25], which allows to efficiently study out-of-equilibrium systems by projecting the time-dependent operators on their equilibrium counterparts, we analyze the field-induced charge carriers in a -wave AM model as well as a metallic RuO bilayer, employing its electronic properties we calculated ab initio.
Our analysis reveals that by changing the linear polarization of the EM pump pulse, it is possible to photo-excite electrons with a specific spin orientation, exploiting the resonance of specific momentum regions of the band dispersion to the pulse frequency and, in particular, the polarization and momentum dependence of the light-matter coupling per spin direction in AMs. We show how the effect can be tuned and optimized adapting the pulse characteristics and demonstrate how the large electronic spin-splittings of AMs in momentum space can be leveraged to generate large densities of pump-induced spin polarization. The pulse polarization control of the spin direction and net magnetization of photo-pumped charge carriers are underlain by the altermagnetic symmetry and thus absent in conventional ferro- or antiferromagnets.
d-wave AM model Hamiltonian—To capture the photo-excitation behavior of an AM induced by a linearly polarized EM pulse, we first consider the simple realization of a -wave AM introduced in Ref. [26]. The model takes the form of a 2D Lieb lattice shown in Figure 1(a), where the anti-parallel magnetic moments (blue and red dots) form two distinct sublattices related by fourfold rotation. The AM character emerges from the two sublattices having distinct local environments due to the presence of further non-magnetic sites [26]. A Kondo-type interaction couples frozen localized spins with itinerant electrons hopping only among the two magnetic sublattices, resulting in a spin-split band structure, see Figure 1(b). For simplicity, we take a parametrization (see the End Matter A) that comes with a clean gap between the AM bands at half filling, see Figure 1(b). As expected, the spin-up, , and spin-down, , bands are related by a rotation in the space [see Figure 1(b)]. We indicate the valence (conduction) band by . The spin-specific band gap, , naturally exhibits the AM symmetry [see Figure 1(c-d)].
Figure 1. (a) Sketch of the -wave AM model and the vector potential of the pump pulse impinging on it. (b) Spin-up (red) and spin-down (blue) electronic bands. The thickness of the superimposed solid lines is proportional to the post-pump electron and hole photo-excited populations, for , , , and . Momenta , for , denote resonant -points. Only host non-negligible post-pump photo-excited populations because of the corresponding noticeable coupling to the pump pulse. (c-d) Band gap over the whole BZ for spin-down and spin-up bands, respectively. The dashed lines mark the positions of resonant -points. (e-f) Light-matter coupling strengths, , over the whole BZ for spin-down and spin-up bands, respectively, and .
Light pulse impinging on a solid—To determine how an ultrafast EM pulse impacts the spin and charge distribution of photo-excited states in an AM, we start from the spin-conserving dipole-gauge Hamiltonian describing the interaction between a classical EM pulse in the dipole approximation and a solid-state lattice system [24, 25, 27]
where is the crystal momentum that is summed over the Brillouin zone (BZ), is the electronic spin, and are sets of quantum numbers identifying orthogonal localized electronic states (e.g., the maximally localized Wannier states), () is the annihilation (creation) operator of an electron in the state with quantum numbers , is the vector potential and is the electric field of the impinging pulse. is the matrix element of the first-quantization (single-particle) equilibrium Hamiltonian and the matrix element of the local dipole element in the reciprocal space.
Mechanism for polarization control of post-pump electron-spin distribution—Before presenting the calculated post-pump, non-equilibrium electron-spin distribution, we first elucidate the mechanism by which the polarization of the light pulse can control and switch the spin orientation of the excited electrons, a property that sets AMs apart from conventional FMs and AFMs. We consider a pump-pulse polarization in the lattice plane parametrized by the angle formed with the axis, as indicated in Figure 1(a). The cross-section for photo-excitation is large when the spin-specific band gap is in resonance with the pump-pulse frequency . The resonant momenta are indicated by dashed lines in Figure 1(c-f) and include, for instance, momenta indicated in Figure 1(b). Here correspond to resonances of spin-up electrons, and to spin-down ones. It is important to note that due to the altermagnetic symmetry, are related to by a rotation. The light-matter coupling strength determines how strongly these states couple to the light field. In an AM, this essential ingredient depends in a non-trivial manner on the spin of the electronic states involved and the polarization of the light pulse. In the following, we will determine the expression of and its dependence on light-pulse characteristics explicitly. The calculated light-matter coupling strength for the d-wave AM model is shown in Figure 1(e,f) for . For this polarization direction, one observes that although are in resonance, these states only couple weakly to the light field so that the post-pump density of spin-down states is low (see Figure 2(d)). However, for the same polarization direction, strongly couple to the light field (Figure 1(f)) and host a large spin-up density after pumping (Figure 2(b)). Because of the altermagnetic symmetry, the rotation of the incoming light-pulse polarization by exchanges the role of spin up and spin down, see Figure 2, giving rise to spin-switching.
Figure 2. The pump pulse polarization controls the spin of the photo-excited charge carriers: shown are the post-pump spin up/down electron populations, , for pump pulse polarization : (a,b) spin-up, (c,d) spin-down, (a,c) , and (b,d) .
Dynamical projective operatorial approach (DPOA)— To determine quantitatively the impact of the pump pulse on the electronic state of the system, we use the DPOA [23, 24, 25], which efficiently and effectively study out-of-equilibrium systems by projecting the time-dependent operators on their equilibrium counterparts, , and moving the solution of the operatorial dynamics to the solution of the equations of motion of the dynamical projection matrices, (see the End Matter B). In principle, this allows to compute any out-of-equilibrium response, such as time-resolved ARPES [24] and transient optical properties [25]. In particular, the electronic band populations can be obtained as , where is the Fermi distribution function [24] (see the End Matter B).
Resulting post-pump electron-spin distribution— The pump pulse that we consider features a vector potential , where is its amplitude, is the FWHM of its Gaussian enve-
lope centered at time , and is its central frequency. The electronic excited population, which is the positive excess post-pump electronic population with respect to the thermal equilibrium one summed over all bands, is denoted by . Given that the main excitation processes in the model are due to one-photon resonances, the dimensionless light-matter coupling is determined from the off-diagonal element of the velocity (first-order) term in the Peierls expansion: (see the End Matter B). exhibits the same symmetry of and under rotation of the polarization. As indicated above, for fixed polarization, is instead very different between spin up and spin down [see Figure 1 (e-f)].
The post-pump electron photo-excited populations, , of the -wave AM model for the four relevant cases obtained by crossing the values of spin (up and down) and polarization ( and ) are shown in Figure 2. The noticeable difference in the coupling strength between spin up and spin down for the chosen pump pulse frequency and polarization (), that is along the dashed lines of Figure 1 (e) and (f), leads to the huge difference between the values of and in Figure 2 (b) and (d), and determines the regions in space where one can find the higher values of : the momentum and spin selectivity by polarization is demonstrated. The spin symmetry of relates Figure 2 (a) and (c), to (b) and (d).
Having established that the pump pulse frequency, , selects the resonant momentum region, and the pump pulse polarization, , is very effective in selecting the spin, we now turn to the effects of the pulse’s amplitude, , and duration, measured through the FWHM, , of its Gaussian envelope [see Figure 3 (a)]. In Figure 3 (b) and (c), we report the post-pump electron photo-excited magnetization per unit cell, , to discuss its dependence on pump pulse FWHM, amplitude, and frequency at fixed polarization ().
Figure 3. (a) Characteristics of the vector potential of the pump pulse. (b-c) Post-pump electron photo-excited magnetization per unit cell, , as a function of the FWHM, , and (b) the square of the pump pulse amplitude, , (c) the pump pulse frequency, , for and the same parameter values of Figure 1.
Within the overall time span of the pump pulse, the electron photo-excited populations at (and close to) resonant -points undergo Rabi-like oscillations with a Rabi frequency, , which is proportional to [28, 29]. Accordingly, the post-pump electron photo-excited populations, , is roughly proportional to , which can be approximated to for small enough pump pulse amplitudes and durations [28]. This explains why, in Figure 3 (b), increases monotonically with the pump pulse FWHM, , only at small enough amplitudes, , while it has a re-entrant behavior on continuously increasing for large enough values of .
On varying the pump pulse frequency, , instead, we change the loci of resonant -points within the BZ and, accordingly, we also explore the landscape of light-matter coupling strengths, [compare to Figure 1 (c-Figure 1f)]. At the largest frequencies, in Figure 3 (c), the region of resonant -points is so small [see Figure 1 (c-Figure 1d)] that the reentrant behavior of on increasing is very difficult to resolve on the 2D map, while it is clearly visible for all other high and intermediate values of the frequency. For smaller and smaller values of the frequency, we observe an interesting new effect: the number of resonant -points with significative values of the coupling strength becomes more and more comparable between the two spin orientations [compare to Figure 1 (c-Figure 1f)] up to an overtaking that reflects in an inversion of the sign of for [the reddish area in Figure 3 (c)]. On increasing in this region of pump pulse frequencies, , given the strong spin dependence of the Rabi frequency, , we scan values of that again tend to compensate between the spin orientations over the whole BZ and increase the value of up to reestablishing its positivity.
**Pumped altermagnetic RuO bilayer—**To demonstrate this phenomenology and the related effects also on the ab initio electronic structure of materials, with all their complexities in terms of band structure and number of active degrees of freedom, we performed a similar analysis on a protype AM material – a RuO bilayer (see the End Matter C). While experiment evidences that bulk RuO is not magnetic [30, 31, 32] [33], altermagnetism can be stabilized in thin films [34, 35]. In real materials, charge excitations decay because of electron-phonon scattering, spontaneous emission, etc. However, the time-scales of such processes are usually of the order of hundreds of femtoseconds [21, 29] [36, 37], while the ultrafast pumping that we consider here occurs over a time-scale of tens of femtoseconds. Accordingly, the post-pump photo-excitations are computed well before such decay mechanisms become relevant.
In Figure 4 (a) and (Figure 4b), the post-pump electron photo-excited populations, , for and , respectively, clearly show that through this protocol—varying the pump pulse polarization—it is possible to excite different spin polarizations selectively both in specific regions of the BZ and overall in the system. This is further confirmed by looking at the post-pump electron photo-excited magnetization per unit cell, , as a function of the pump pulse characteristics. In Figure 4 (c), despite the much larger complexity of the system under analysis, we recover the same qualitative behavior we have found for the -wave AM model [compare to Figure 3 (b)] confirming the robustness of the phenomenology.
Figure 4. Post-pump electron photo-excited populations, , for the bilayer RuO for (a) and (b) . The used pump pulse parameters are V fs/nm, PHz, and fs. The temperature is fixed at about 290 K (i.e., 25 meV). (c) Post-pump electron photo-excited magnetization per unit cell, , for as a function of the pump pulse FWHM, , and the square of the pump pulse amplitude, , with PHz and (d) the pump pulse frequency, , and for , 20 and 28 fs with V fs/nm.
In Figure 4 (d), we see that can be easily tuned in absolute value and even in sign, on varying both the pump pulse frequency, , and the FWHM, , of the pump pulse. The complexity of real materials calls for different regions of the BZ for each spin orientation to become active at the same time, in terms of being single or multi-photon resonant and more or less coupled to the applied EM pump pulse. It is worth noting that, in real materials, we can have multi-photon resonances, and not necessarily just single-photon ones, and the former can become even dominant with respect to the latter [23].
The complexity of real materials does not hinder the phenomenology, which we proved to be very robust even in this case. Actually, such complexity allows for greater tunability and switchability, the possibility of designing and engineering altermagnetic materials for specific needs, and paves the way for an effective ultrafast optical control of AMs, with all possible future potential technological applications.
Conclusions and Perspectives—The phenomenology and the effects discovered and demonstrated can be used (i) to identify altermagnetic materials via pump-probe Kerr and Faraday spectroscopies and/or spin- and time-resolved ARPES, (ii) to probe the spin polarization of the band structure of AMs in energy and momentum space, (iii) to generate photo-excited holes and electrons with a specific spin orientation that could be used to obtain bias-induced spin-polarized currents, and (iv) to devise an optical switch of the net spin polarization of photo-excited holes and electrons by tuning the pump pulse characteristics.
Acknowledgments—AE and AA acknowledge financial support from PNRR MUR project PE0000023-NQSTI-TOPQIN. JvdB and OJ acknowledge financial support from the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) through the Sonderforschungsbereich SFB 1143, grant No. YE 232/2-1, and under Germany’s Excellence Strategy through the Würzburg-Dresden Cluster of Excellence on Complexity and Topology in Quantum Matter – ct.qmat (EXC 2147, project-ids 390858490 and 392019).
References
End Matter
The d-wave AM model
In the -wave AM model [26], the Hamiltonian is parameterized by the Kondo interaction strength, , the staggered magnetization vector, , a nearest-neighbor hopping, , and anisotropic next-nearest-neighbor hoppings, and , over two sublattices A and B. Then, its matrix elements read as [26],
where , , , and are the Pauli matrices, and () corresponds to the spin up (down). We set , which is the staggered magnetization in the direction. Moreover, we set , and . In this model, .
Dynamical Projective Operatorial Approach (DPOA)
In principle, for a general time-dependent Hamiltonian , it is always possible to find some sets of operators, , that display a closed hierarchy of equations of motion: , where is the matrix product in the space of the operators belonging to a specific set , while is known as the time-dependent energy matrix [24] [38]. The DPOA [23, 24, 25] exploits this occurrence to efficiently and effectively study out-of-equilibrium systems by projecting the time-dependent operators on their equilibrium counterparts () and moving the solution of the operatorial dynamics to the solution of the equations of motion of the dynamical projection matrices : . For a quadratic Hamiltonian of a solid-state lattice system, the operators reduce to the electronic annihilation operators.
Usually, it is convenient to work in a basis that diagonalizes the equilibrium Hamiltonian: in Eq. (1). The transformation to such a basis, which we denote by the index , is performed by a unitary matrix, . The electronic annihilation operator in this basis reads as . The corresponding eigenvalues are denoted by , and are the bands of the system. One fundamental ingredient in the DPOA framework is the use of the so-called Peierls expansion, which is a Taylor expansion for computing very efficiently and in Eq. (1) [24]: , where and is the polarization of .
Further details on the RuO bilayer
While there is a mounting experimental evidence that bulk RuO is not magnetic [30, 31, 32] [33], altermagnetism can be stabilized in thin films [34, 35]. Here, we consider a free-standing bilayer of RuO, which we use a prototype system with an effective -wave AM model that mimics the complexity of real materials. To construct an altermagnetic bilayer, we start from the bulk crystal structure and make a slab with two formula units oriented along the -axis of the bulk system. Opposite-spin Ru ions are related by fourfold roto-inversion symmetry, and the system can be represented as a supercell described within the space group (35) with Å. For this structure, we perform density-functional calculations in GGA+ [39, 40], as implemented in the Wien2k code [41, 42]. We use the full localized limit for the double counting correction and parameters and . For the BZ integrations, we use a -point mesh having 2000 -points in the irreducible BZ, along with a tetrahedron method.
The calculation of the post-pump excitations exploits a Wannier model based on Ru and O orbitals, which was obtained with the Wannier90 code [43]. The model successfully reproduces the DFT band structure in the energy range around the Fermi energy.
References
- [1]L. Šmejkal, R. González-Hernández, T. Jungwirth, and J. Sinova, Science Advances 6, eaaz8809 (2020).
- [2]L. Šmejkal, J. Sinova, and T. Jungwirth, Phys. Rev. X 12, 031042 (2022).
- [3]L. Šmejkal, J. Sinova, and T. Jungwirth, Physical Review X 12, 040501 (2022).
- [4]M. Naka, S. Hayami, H. Kusunose, Y. Yanagi, Y. Motome, and H. Seo, Nature Communications 10, 4305 (2019), arXiv:1902.02506 [cond-mat.str-el].DOI
- [5]L.-D. Yuan, Z. Wang, J.-W. Luo, E. I. Rashba, and A. Zunger, Phys. Rev. B 102, 014422 (2020).
- [6]M. Naka, Y. Motome, and H. Seo, Phys. Rev. B 103, 125114 (2021).
- [7]Y. Guo, H. Liu, O. Janson, I. C. Fulga, J. van den Brink, and J. I. Facio, Materials Today Physics 32, 100991 (2023).
- [8]P. A. McClarty and J. G. Rau, Physical Review Letters 132, 176702 (2024).
- [9]T. Sato, S. Haddad, I. C. Fulga, F. F. Assaad, and J. van den Brink, Phys. Rev. Lett. 133, 086503 (2024).
- [10]M. Trama, I. Gaiardoni, C. Guarcellо, J. I. Facio, A. Maiellaro, F. Romeo, R. Citro, and J. van den Brink, “Non-linear anomalous edelstein response at altermagnetic interfaces,” (2024), arXiv:2410.18036 [cond-mat.mtrl-sci].
- [11]M. Hu, O. Janson, C. Felser, P. McClarty, J. van den Brink, and M. G. Vergniory, “Spin hall and edelstein effects in novel chiral noncollinear altermagnets,” (2024), arXiv:2410.17993 [cond-mat.mtrl-sci].
- [12]P. M. Consoli, M. Fornville, and M. Vojta, Physical Review B 104, 064422 (2021).
- [13]T. Aoyama and K. Ohgushi, Phys. Rev. Mater. 8, L041402 (2024).
- [14]K. V. Yershov, V. P. Kravchuk, M. Daghofer, and J. van den Brink, Phys. Rev. B 110, 144421 (2024).
- [15]K. V. Yershov, O. Gomonay, J. Sinova, J. van den Brink, and V. P. Kravchuk, “Curvature induced magnetization of altermagnetic films,” (2024), arXiv:2406.05103 [cond-mat.str-el].
- [16]M. Vila, V. Sunko, and J. E. Moore, “Orbital-spin locking and its optical signatures in altermagnets,” (2024), arXiv:2410.23513 [cond-mat.mtrl-sci].
- [17]T. Brabec and F. Krausz, Reviews of Modern Physics 72, 545 (2000).
- [18]F. Schmitt, P. S. Kirchmann, U. Bovensiepen, R. G. Moore, L. Rettig, M. Krenz, J.-H. Chu, N. Ru, L. Perfetti, D. Lu, et al., Science 321, 1649 (2008).
- [19]F. Krausz and M. Ivanov, Reviews of modern physics 81, 163 (2009).
- [20]F. Krausz and M. I. Stockman, Nature Photonics 8, 205 (2014).
- [21]M. Zürch, H.-T. Chang, L. J. Borja, P. M. Kraus, S. K. Cushing, A. Gandman, C. J. Kaplan, M. H. Oh, J. S. Prell, D. Prendergast, C. D. Pemmaraju, D. M. Neumark, and S. R. Leone, Nature Communications 8, 15734 (2017).
- [22]R. Borrego-Varillas, M. Lucchini, and M. Nisoli, Reports on Progress in Physics 85, 066401 (2022).
- [23]G. Inzani, L. Adamska, A. Eskandari-asl, N. D. Palo, G. L. Dolso, B. Moio, L. J. D’Onofrio, A. Lamperti, A. Molle, R. Borrego-Varillas, M. Nisoli, S. Pittalis, C. A. Rozzi, A. Avella, and M. Lucchini, Nature Photonics 17, 1059 (2023).
- [24]A. Eskandari-asl and A. Avella, Physical Review B 110, 094309 (2024).
- [25]A. Eskandari-asl and A. Avella, Physical Review A 110, 043520 (2024).
- [26]D. S. Antonenko, R. M. Fernandes, and J. W. F. Venderbos, “Mirror chern bands and weyl nodal loops in altermagnets,” (2024), arXiv:2402.10201 [cond-mat.mes-hall].
- [27]M. Schüler, J. A. Marks, Y. Murakami, C. Jia, and T. P. Devereaux, Physical Review B 103, 155409 (2021).
- [28]A. Eskandari-asl and A. Avella, in Advances in Ultrafast Condensed Phase Physics IV, Vol. 192 (SPIE, 2024) pp. 69–73.
- [29]G. Inzani, L. Adamska, A. Eskandari-asl, N. Di Palo, G. L. Dolso, B. Moio, L. J. D’Onofrio, A. Lamperti, A. Molle, R. Borrego-Varillas, M. Nisoli, et al., Nature Photonics 17, 1059 (2023).
- [30]M. Hirai, H. Okabe, A. Koda, R. Kadono, T. Muroi, D. Hirai, and Z. Hiroi, Phys. Rev. Lett. 132, 166702 (2024).
- [31]J. Liu, J. Zhan, T. Li, J. Liu, S. Cheng, Y. Shi, L. Deng, M. Zhang, C. Li, J. Ding, Q. Jiang, M. Ye, Z. Liu, Z. Jiang, S. Wang, Q. Li, Y. Xie, Y. Wang, S. Qiao, J. Wen, Y. Sun, and D. Shen, Phys. Rev. Lett. 133, 176401 (2024).
- [32]L. Kiefer, F. Wirth, A. Bertin, P. Becker, L. Bo-
