WorksPhysics and Astronomy
Crystal Symmetry Selected Pure Spin Photocurrent in Altermagnetic Insulators
Abstract
The generation of time-reversal-odd spin-current in metallic altermagnets has attracted considerable interest in spintronics. However, producing pure spin-current in insulating materials remains both challenging and desirable, as insulating states are frequently found in antiferromagnets. Nonlinear photogalvanic effects offer a promising method for generating spin-current in insulators. We here revealed that spin and charge photocurrents in altermagnets are protected by spin point group symmetry. Unlike the photocurrents in parity-time symmetric materials, where spin-orbit coupling (SOC) induces a significant charge current, the spin-current in altermagnets can exist as a pure spin current along specific crystal directions regardless of SOC. We applied our predictions using first-principles calculations to several distinct materials, including wurtzite MnTe and multiferroic BiFeO3. Additionally, we elucidated the previously overlooked linear-inject-current mechanism in BiFeO3 induced by SOC, which may account for the enhanced bulk photovotaic effect in multiferroics.
Introduction
Generating spin-polarized current, particularly pure spin current, in solid-state systems has received significant attention [1, 2]. A promising approach to generate pure spin-current is via the spin Hall effect in metals [3, 4, 5] or the quantum spin Hall effect in topological insulators[6, 7]. These mechanisms typically do not require magnetic order but are at the expense of relying on spin-orbit coupling (SOC). Recently, the altermagnetic phase, characterized by alternating spin-splitting bands resulting from real-space rotation transformation, has been revealed [8, 9, 10]. Its significant spin splitting induced by strong non-relativistic exchange coupling, along with its conservation of spin momentum are viewed as promising developments for spintronics [8, 9, 10, 11, 12] and possible applications [13, 14, 15, 16]. For example, the broken time-reversal symmetry () in altermagnets enables the generation of -odd spin current in metals[17, 18] such as RuO, regardless of SOC. Generating spin current in insulators via such a DC bias is impossible; however, spin-current in altermagnetic insulators is highly desirable, as insulating states constitute a large portion of antiferromagnets [1, 19].
Nonlinear photogalvanic effects are viewed as a promising approach to generate spin-polarized current in non-centrosymmetric insulators[20, 21, 22, 23, 24, 25, 26, 27]. Notably, the two mechanisms, shift current and inject current[28, 29], exhibit different parity under time-reversal () or parity-time () symmetry[24, 25, 30, 31, 32, 33]. This allows them to produce spin or charge current, respectively, making them suitable for manipulation through optical linear or circular polarization. For example, the linear-shift-current mechanism (referring to shift current induced by linearly polarized light) would generate spin current in -symmetric antiferromagnets (-AFMs)[23], while the circular-inject-current mechanism produces spin current[26]. However, both linear-inject-current and
circular-shift-current mechanisms would generate charge current in -AFMs with SOC[34, 35], rendering pure spin current impossible. In contrast to -AFM, where the spin-down and spin-up sublattices are connected by symmetry, they are linked by real-space rotation transformation in altermagnets[36]. This provides a solid platform for manipulating both spin and charge currents through its crystal symmetry using light.
In this work, we reveal the crystal symmetry-dependent photo-driven spin and charge currents in altermagnets. Unlike -AFM, the spin-up and spin-down sublattices in altermagnets are connected by real-space rotational transformation, allowing spin-up and spin-down electrons to exhibit covariant behavior in shift-current and inject-current mechanisms. Through symmetry analysis, we demonstrate that both mechanisms can generate spin-current in specific directions, leading to pure spin-current regardless of SOC. We validate these predictions through first-principles calculations of two experimentally accessible altermagnets: wurtzite MnTe and multiferroic BiFeO. The spin currents, generated by these two mechanisms, are significant and codirectional across a wide range of photonic energies.
Pure spin-current independent of SOC— In the case of coherent light illumination, the primary mechanisms responsible for a second-order DC photocurrent are the shift current and inject current [28, 29, 37]. Here, represents the electrical field of incoming light with frequency , with indices and denote the polarization directions of the light, and specifying the current direction. The shift current primarily characterizes the polarization variance between the conduction and valence bands [28, 30]:
where denotes the difference in occupation numbers between two bands. is the velocity matrix element, and defines the gauge-independent ‘generalized derivatives’ of velocity matrix element. stands for the Berry connection of band , while denotes the shift vector. Conversely, the inject current shows a linear dependence on the carrier lifetime ()[28]
When examining the photo-driven spin-current, we substitute the velocity matrix in the direction with , where with denoting the spin component along the -direction. The “” in “” signifies linearly polarized light, while “” represents circularly polarized light for both shift current and inject current. This implies that the linearly polarized light and circularly polarized light cases depend on the real and imaginary parts of , respectively. The distinct parities of the real and imaginary parts of under various symmetry operators, such as and , are pivotal in determining the charge or spin current in photogalvanic effects. Similar scenarios were found in the -odd or -even spin Hall effects of antiferromagnetic metals[17, 38].
In antiferromagnets with symmetry, the spin-up and spin-down electrons with energy degeneracy at the same -point are connected by the symmetry in the reciprocal space. When applying the symmetry on spin ()-dependent velocity matrix and shift vector, we obtained and . As a result, the real and imaginary components of exhibit -even and -odd characteristics for opposite spins, respectively. Following Eqs. (1) and (2), for linearly polarized light, which is associated with the -even real component, the shift current and inject current generate spin-current and charge-current [23, 34, 35], respectively (See Table 1). However, under circularly polarized light, this scenario reverses entirely due to the -odd characteristics of the imaginary component[26].
| Linearly | Circularly | ||||
| Shift | Inject | Shift | Inject | ||
| -AFM | non-SOC | spin | 0 | 0 | spin |
| SOC | spin | charge | charge | spin | |
| Altermagnet | non-SOC | spin ( or )<br>charge() | 0 | 0 | spin ( or )<br>charge() |
| SOC | spin ( or )<br>charge() | spin ( or )<br>charge() | spin ( or )<br>charge() | spin ( or )<br>charge() |
Table 1. The spatial separation of spin and charge currents in altermagnets, contrasts with -antiferromagnets, where charge current is always present due to SOC. This is exemplified by scenarios where the spin-up and spin-down sub-lattices in altermagnet are connected by rotational symmetry along the -direction, e.g. spin point group, with light polarization in the plane. ‘Linearly’ and ‘Circularly’ refer to linearly and circularly polarized light, respectively.
Note that spin-rotation symmetry is preserved in a system without SOC [36, 39]. Here the is a rotation around an axis perpendicular to the spins, combined with the spin-space inversion; the is a time-reversal operator in the real space which flips the sign of the crystal momentum. Applying the spin-rotation transformation on spin ()-dependent bands and velocity matrix, we obtained and . Consequently, the real and imaginary parts of and group velocity are -even and -odd respectively, causing a zero charge-current from the linear-inject-current mechanism according to (2). Similarly, the imaginary component of and shift vector are -odd and -even respectively, causing a zero charge-current from circular-shift-current mechanism according to (1). Thus, as depicted in Figure 1(a), the linear-shift-current or the circular-inject-current would generate pure spin-current without any charge-current when SOC is absent.

Figure 1. The schematic scenarios illustrate the linearly polarized light driven spin-current (a) and the charge-current resulting from SOC (b) in -AFMs, which disrupt the pure spin current. The crystal direction-dependent spin- or charge-current in altermagnetic insulators for both shift-current (c) and inject-current (d). The Brillouin zone of an altermagnet with a spin point group, featuring two spin-splitting off-nodal planes (e) and (f).
The situation is entirely different for altermagnets, as illustrated in Figure 1b. The spin-up and spin-down electrons, which are energy-degenerate at different -point, are connected by spin group transformation instead of the symmetry as in -AFMs[36]. Here, the transformation denotes a rotation in the spin-only space, while transformation encompasses solely real-space proper or improper rotations [36]. In this scenario, the linear-shift-current mechanism can simultaneously generate spin and charge currents along different axes, as illustrated in Figure 1(b). This situation similarly applies to the circular-inject-current mechanism. If SOC is not considered, both the linear-inject-current and circular-shift-current result in zero charge or spin current, similar to the case in -AFMs.
Without loss of generality, we discuss an altermagnet with a spin point group , where the spin-up and spin-down sublattices transform via . This transformation classifies it as anisotropic -wave altermagnets [36]. As illustrated in Figure 1c and Figure 1d, the Brillouin zone is partitioned into four regions by the gray-shaded nodal planes. To demonstrate the influence of symmetry, we depict the spin polarization of one band, with spin-up and spin-down regions shown in red and blue, respectively. For the spin-splitting on the off-nodal plane , the pattern exhibits inversion symmetry in -space owing to the transformation , which persists in systems without SOC, regardless of crystal inversion. While on the off-nodal plane , which is parallel to the spin-degenerate plane but with , the spin polarization demonstrates a two-fold rotation symmetry governed by spin-point group transformation .
When applying the spin-rotation transformation on spin ()-dependent velocity matrix, we obtained . This results in zero charge-current governed by the linear-inject-current and circular-shift-current mechanisms, similar to that in -AFMs. However, applying the transformation on the velocity matrix results in distinct parities: even for the - and -directions, but odd for the -direction. Specifically, we get: and . The shift vector undergoes similar transformation relations under . These properties suggest that spin-up and spin-down electrons, connected by , will travel in opposite directions along the or axes but will move in the same direction along the axis. Consequently, the effective transportation of carriers in the or direction forms the spin-current, while the charge-current flows along the direction.
When considering SOC, the spin point group transitions to the magnetic point group in collinear antiferromagnets, allowing for a non-zero linear-inject current (See Table 1). Under rotation symmetry , the spin-dependent velocity matrix is translated to , resulting in -even real component of and -odd group velocity. Consequently, a nonzero spin current is generated along the or direction and a charge current along the direction through the linear-inject-current mechanism. Thus, a pure spin current can be induced along the or direction. This pattern holds for scenarios involving circularly polarized light with polarization in the plane.
Wurtzite MnTe—The -MnTe with its NiAs-type structure possessing inversion symmetry [40, 41, 42], lacks the capability for generating second-order photo-driven spin currents. In contrast, the wurtzite form of MnTe (spin point group ), depicted in Figure 2a, breaks inversion symmetry, leading to a notable photogalvanic effect. Despite -MnTe being the energetically favorable structure [43], the wurtzite configuration remains metastable and has been successfully synthesized and maintained at room temperature in experiments, demonstrating a moderate band-gap suitable for solar cell applications [44, 45].

Figure 2. (a) The crystal structure of wurtzite MnTe, (b) the top view of two opposite-spin sublattices (Mn and Mn), which are connected through spin rotation combined with sixfold crystal rotation and half-unit cell translation () along the axis. (c) the Brillouin zone and the off-nodal plane with , (d) the off-nodal plane. The spin-current (e) and charge-current (f) conductivities without SOC originated from the linear-shift-current mechanism. The corresponding spin photoconductivity distribution on the plane (g) and the plane (h) without SOC for 3.5 eV photons.
We discovered that the wurtzite structure exhibits characteristics of a -wave altermagnet, similar to the -phase. However, unlike the -phase, the two opposite-spin sublattices (Mn and Mn) in the wurtzite structure transform through (Figure 2(b)), losing the symmetry that links the two sublattices in the -phase (See supplement material [46? ? ? ? – 50]). In Figure 2c and 2d, the spin-degenerate nodal planes within the Brillouin zone are shaded in grey. The spin-degenerate plane is protected by combined with , and the plane is protected by combined with ( is the spin-space identity). The pattern on the off-nodal plane exhibits inversion symmetry owing to the spin-rotation operator , which persists in systems without SOC. While for bands on the off-nodal plane , which is parallel to but with , the spin-splitting demonstrates a three-fold pattern governed by the crystal rotation operator , and the spin-up and spin-down is connected by , leading to the pattern shown in Figure 2d.
If SOC is not considered, our first-principles calculations confirm that the linear-inject-current and circular-shift-current mechanisms contribute neither to spin-current nor charge-current. This is universally observed in -AFMs and altermagnets, as listed in Table 1. However, as previously discussed, the linear-shift-current and circular-inject-current mechanisms can contribute to nonzero spin or charge currents. These currents depend on the spin point group , which has a transformation .
To simplify the analysis, we apply the transformation to matrix . The resulting velocity matrices exhibit distinct parities: even for the - and -directions, but odd for the -direction. These characteristics indicate that the transport of carriers in the or direction results in spin current, while the charge current flows along the direction.
The currents induced by the linear-shift-current mechanism are illustrated in Figure 2e and Figure 2f, where the spin-current is nonzero, and the charge-current is nonzero as well. The and components spin-current are zero due to the -directional easy-axis of the Néel vector. Interestingly, the spin and charge currents in the -direction are zero due to mirror symmetry . Figure 2g and Figure 2h illustrate the parities of shift-current conductivity under and operators, respectively. Both demonstrate an even behavior for the spin-conductivity tensor , leading to a non-zero spin-current (See circular-inject-current in supplementary [24]). When SOC is considered, the inject-current and shift-current mechanisms still display the same behavior under the transformations of spin point-group symmetry, resulting in pure spin currents in a specific direction (see supplement information [24]).
Multiferroic BiFeO: Bismuth ferrite (BFO) exhibits a ferroelectric phase transition Curie temperature () of 1083–1103 and a G-type antiferromagnetic phase transition Néel temperature () of 625–643 , both significantly surpassing room temperature[50, 51, 52]. This exceptional and make BFO the leading candidate and widely recognized multiferroic material. The polarization direction corresponds to the [111] axis of the rhombohedra primitive cell, aligning with the -axis of the hexagonal structure, as illustrated in Figure 3a. Additionally, the easy magnetization plane is perpendicular to the ferroelectric polarization axes. The two opposite-spin sublattices Fe and Fe are related by the transformation (Figure 3b). Additionally, the symmetry places BFO in the spin-point group , classifying it as a g-wave altermagnet (see supplement information [24]).

Figure 3. (a) The crystal structure of multiferroic BiFeO (b) the top view of two opposite-spin sublattices (Fe and Fe), which are related by spin rotation combined with mirror symmetry and half-unit cell translation (). The spin-current (c) and charge-current (d) conductivities arise from the linear-shift-current mechanism without SOC. The effective Hubbard is set to 5 eV because the calculated imaginary permittivity best matches the experimental results for energies near the band gap.
Our first-principles calculations confirm that only the linear-shift-current and circular-inject-current mechanisms contribute to the spin-current or charge-current in the absence of SOC. Figure 3c and 3d illustrate the spin-current and charge-current induced by the linear-shift-current mechanism, respectively. For instance, when light is linearly polarized along the -direction, spin-up and spin-down carriers effectively generate a -direction spin-current, while a or directions charge-current due to the transformation .
In BFO, SOC plays a crucial role in light-induced spin currents and charge currents. Table 1 outlines the SOC enables linear-inject-current and circular-shift-current mechanisms, as also confirmed in BFO through first-principles calculations. For example, Figure 4a and 4b show the currents from the linear-shift-current mechanism. In contrast to Figure 3c and Figure 3d, SOC can reverse the direction of the spin-current (e.g. ) by modifying the Berry connection of the bands. However, whether conductivity components are nonzero or not remains unaffected, as they are determined by the crystal symmetry.

Figure 4. The photo-driven spin- and charge-current of multiferroic BiFeO with SOC using eV. The spin-current (a) and charge-current (b) via the linear-shift-current mechanism. And the spin-current (c) and charge-current (d) by the linear-inject-current mechanism.
Figure 4c and Figure 4d illustrate spin-current and charge-current induced by linear-inject-current with SOC. Interestingly, the spin-current susceptibilities induced by both shift-current and inject-current are in the same direction across a wide range of photonic energies, potentially resulting in a pure spin current, such as and . Interestingly, the charge current contributed by the inject-current is of the same order as that induced by the shift-current, a factor overlooked in previous studies. This additional contribution might explain the discrepancies between calculations and experimental observations [53, 54].
In summary, we have demonstrated that the photo-driven spin and charge currents in altermagnets directed separately due to the spin point group. This behavior is fundamentally different from the symmetry in -AFMs or -symmetry in nonmagnetic materials, making altermagnets uniquely capable of generating pure spin-current. Furthermore, these crystal symmetry-selected pure spin-current can be completely switched based on the type of light used, whether linearly or circularly polarized, offering flexibility in tuning the spin current. First-principles calculations further validate these findings in experimentally obtained materials, such as wurtzite MnTe and BiFeO, which could be beneficial for both spintronic devices and Néel vector detection.
Acknowledgments
This work is supported by the National Natural Science Foundation of China Grant No. 12204035.
References
- [1]V. Baltz, A. Manchon, M. Tsoi, T. Moriyama, T. Ono, and Y. Tserkovnyak, Reviews of Modern Physics 90, 015005 (2018).
- [2]J. Sinova, S. O. Valenzuela, J. Wunderlich, C. Back, and T. Jungwirth, Reviews of Modern Physics 87, 1213 (2015).
- [3]M. Dyakonov and V. Perel, Physics Letters A 35, 459 (1971).
- [4]S. Murakami, N. Nagaosa, and S.-C. Zhang, Science 301, 1348 (2003).
- [5]J. Sinova, D. Culcer, Q. Niu, N. A. Sinitsyn, T. Jungwirth, and A. H. MacDonald, Physical Review Letters 92, 126603 (2004).
- [6]C. L. Kane and E. J. Mele, Physical Review Letters 95, 226801 (2005).
- [7]B. A. Bernevig and S.-C. Zhang, Physical Review Letters 96, 106802 (2006).
- [8]L. Šmejkal, R. González-Hernández, T. Jungwirth, and J. Sinova, Science Advances 6, eaaz8809 (2020).
- [9]S. Hayami, Y. Yanagi, and H. Kusunose, Journal of the Physical Society of Japan 88, 123702 (2019).
- [10]L.-D. Yuan, Z. Wang, J.-W. Luo, E. I. Rashba, and A. Zunger, Physical Review B 102, 014422 (2020).
- [11]I. I. Mazin, K. Koepernik, M. D. Johannes, R. González-Hernández, and L. Šmejkal, Proceedings of the National Academy of Sciences 118, e2108924118 (2021).
- [12]H. Bai, Y. Zhang, Y. Zhou, P. Chen, C. Wan, L. Han, W. Zhu, S. Liang, Y. Su, X. Han, F. Pan, and C. Song, Physical Review Letters 130, 216701 (2023).
- [13]J. A. Ouassou, A. Brataas, and J. Linder, Physical Review Letters 131, 076003 (2023).
- [14]A. Hariki, A. Dal Din, O. J. Amin, T. Yamaguchi, A. Badura, D. Kriegner, K. W. Edmonds, R. P. Campion, P. Wadley, D. Backes, L. S. I. Veiga, S. S. Dhesi, G. Springholz, L. Šmejkal, K. Výborný, T. Jungwirth, and J. Kuneš, Physical Review Letters 132, 176701 (2024).
- [15]R. A. Leenders, D. Afanasiev, A. V. Kimel, and R. V. Mikhaylovskiy, Nature 630, 335 (2024).
- [16]T. Aoyama and K. Ohgushi, Physical Review Materials 8, L041402 (2024).
- [17]R. González-Hernández, L. Šmejkal, K. Výborný, Y. Yahagi, J. Sinova, T. Jungwirth, and J. Železný, Physical Review Letters 126, 127701 (2021).
- [18]H.-Y. Ma, M. Hu, N. Li, J. Liu, W. Yao, J.-F. Jia, and J. Liu, Nature Communications 12, 2846 (2021).
- [19]Z.-F. Gao, S. Qu, B. Zeng, Y. Liu, J.-R. Wen, H. Sun, P.-J. Guo, and Z.-Y. Lu, arXiv , 2311.04418 (2023).
- [20]R. D. R. Bhat, F. Nastos, A. Najmaie, and J. E. Sipe, Physical Review Letters 94, 096603 (2005).
- [21]S. D. Ganichev, E. L. Ivchenko, V. V. Bel’kov, S. A. Tarasenko, M. Sollinger, D. Weiss, W. Wegscheider, and W. Prettl, Nature 417, 153 (2002).
- [22]B. Zhou and S.-Q. Shen, Physical Review B 75, 045339 (2007).
- [23]S. M. Young, F. Zheng, and A. M. Rappe, Physical Review Letters 110, 057201 (2013).
- [24]R. Fei, X. Lu, and L. Yang, arXiv , 2006.10690 (2020).
- [25]H. Xu, H. Wang, J. Zhou, and J. Li, Nature Communications 12, 4330 (2021).
- [26]R. Fei, W. Song, L. Pusey-Nazzaro, and L. Yang, Physical Review Letters 127, 207402 (2021).
- [27]T. Song, E. Anderson, M.-W.-Y. Tu, K. Seyler, T. Taniguchi, K. Watanabe, M. A. McGuire, X. Li, T. Cao, D. Xiao, W. Yao, and X. Xu, Science Advances 7, eabg8094 (2021).
- [28]J. E. Sipe and A. I. Shkrebtii, Physical Review B 61, 5337 (2000).
- [29]R. von Baltz and W. Kraut, Physical Review B 23, 5590 (1981).
- [30]J. Ahn, G.-Y. Guo, and N. Nagaosa, Physical Review X 10, 041041 (2020).
- [31]H. Watanabe and Y. Yanase, Physical Review X 11, 011001 (2021).
- [32]H. Wang and X. Qian, Science Advances 5, eaav9743 (2019).
- [33]N. Nagaosa and Y. Yanase, Annual Review of Condensed Matter Physics 15, 63 (2024).
- [34]Y. Zhang, T. Holder, H. Ishizuka, F. de Juan, N. Nagaosa, C. Felser, and B. Yan, Nature Communications 10, 3783 (2019).
- [35]R. Fei, W. Song, and L. Yang, Physical Review B 102, 035440 (2020).
- [36]L. Šmejkal, J. Sinova, and T. Jungwirth, Physical Review X 12, 031042 (2022).
- [37]W. Kraut and R. V. Baltz, Physical Review B 19, 1548 (1979).
- [38]F. Freimuth, S. Blügel, and Y. Mokrousov, Phys. Rev. B 90, 174423 (2014).
- [39]P. Liu, J. Li, J. Han, X. Wan, and Q. Liu, Physical Review X 12, 021016 (2022).
- [40]S. Lee, S. Lee, S. Jung, J. Jung, D. Kim, Y. Lee, B. Seok, J. Kim, B. G. Park, L. Šmejkal, C.-J. Kang, and C. Kim, Physical Review Letters 132, 036702 (2024).
- [41]J. Krempaský, L. Šmejkal, S. W. D’Souza, M. Hajlaoui, G. Springholz, K. Uhlířová, F. Alarab, P. C. Constantinou, V. Strocov, D. Usanov, W. R. Pudelko, R. González-Hernández, A. Birk Hellenes, Z. Jansa, H. Reichlová, Z. Šobáň, R. D. Gonzalez Betancourt, P. Wadley, J. Sinova, D. Kriegner, J. Minář, J. H. Dil, and T. Jungwirth, Nature 626, 517 (2024).
- [42]T. Osumi, S. Souma, T. Aoyama, K. Yamauchi, A. Honma, K. Nakayama, T. Takahashi, K. Ohgushi, and T. Sato, Physical Review B 109, 115102 (2024).
- [43]D. Kriegner, H. Reichlova, J. Grenzer, W. Schmidt, E. Ressouche, J. Godinho, T. Wagner, S. Y. Martin, A. B. Shick, V. V. Volobuеv, G. Springholz, V. Holý, J. Wunderlich, T. Jungwirth, and K. Výborný, Physical Review B 96, 241418 (2017).
- [44]S. Siol, Y. Han, J. Mangum, P. Schulz, A. M. Holder, T. R. Klein, M. F. A. M. van Hest, B. Gorman, and A. Zakutayev, Journal of Materials Chemistry C 6, 6297 (2018).
- [45]S. Mori, S. Hatayama, Y. Shuang, D. Ando, and Y. Sutou, Nature Communications 11, 85 (2020).
- [50]S. V. Kiselev, R. P. Ozerov, and G. S. Zhdanov, Soviet physics. Doklady 7, 742 (1963).
- [51]J. R. Teague, R. Gerson, and W. James, Solid State Communications 8, 1073 (1970).
- [52]P. Fischer, M. Polomska, I. Sosnowska, and M. Szymanski, Journal of Physics C: Solid State Physics 13, 1931 (1980).
- [53]S. M. Young, F. Zheng, and A. M. Rappe, Physical Review Letters 109, 236601 (2012).
- [54]S. Y. Yang, J. Seidel, S. J. Byrnes, P. Shafer, C.-H. Yang, M. D. Rossell, P. Yu, Y.-H. Chu, J. F. Scott, J. W. Ager, L. W. Martin, and R. Ramesh, Nature Nanotechnology 5, 143 (2010).