betaFeedback

WorksPhysics and Astronomy

Magnetic Weyl semimetals as a source of circularly polarized THz radiation

Abstract

We propose to use the electromagnetic radiation induced by a few MeV electron beam in magnetic Weyl semimetals as a source of the circularly polarized photons in the THz frequency range.

Introduction. — A fast charged particle traversing a chiral medium emits, by means of the chiral anomaly, a peculiar form of circularly polarized electromagnetic radiation known as the Chiral Cherenkov radiation [1, 2, 3]. It appears as a result of anomalous response of the chiral medium to the electromagnetic field of a fast charge. Therefore, it should appear in a variety of systems ranging from quark gluon plasma to Weyl and Dirac semimetals [4]. In this letter we argue that Chiral Cherenkov radiation in Weyl semimetals is a useful source of polarized electromagnetic radiation in the practically important THz frequency range.

Photon dispersion in Weyl semimetals.— An external electromagnetic field in Weyl semimetals induces the anomalous Hall current

jAH=b×E,(1)\mathbf{j}_{\mathrm{AH}}=\mathbf{b}\times\mathbf{E}, \tag*{(1)}

where b=(α/π)Δ\mathbf{b}=(\alpha/\pi)\boldsymbol{\Delta}, with Δ\boldsymbol{\Delta} being the separation of Weyl nodes in momentum space and α\alpha the fine structure constant. A typical Weyl semimetal has more than one pair of nodes in which case Δ\boldsymbol{\Delta} stands for an effective displacement. It was suggested in [5] that the total displacement is the vector sum of all displacements: Δ=∑iΔi\boldsymbol{\Delta}=\sum_i\boldsymbol{\Delta}_i. In this case Δ\boldsymbol{\Delta} is finite only in magnetic Weyl semimetals but vanishes in non-magnetic ones. For example, in Co3Sn2S2\mathrm{Co}_3\mathrm{Sn}_2\mathrm{S}_2 b=3.8 eV\mathbf{b}=3.8\ \mathrm{eV} [5, 6].

In the presence of the current (1) the dielectric tensor takes form ϵij=ϵδij−iϵijkbk/ω\epsilon_{ij}=\epsilon\delta_{ij}-i\epsilon_{ijk}b_k/\omega, where ϵ≈1/(1+ωp2/ω2)\epsilon\approx1/(1+\omega_p^2/\omega^2) for ω≫ωp\omega\gg\omega_p. The photon dispersion relation is a solution of the Fresnel equation

∣kikj−k2δij+ω2ϵij∣=0(2)\left|k_i k_j-k^2\delta_{ij}+\omega^2\epsilon_{ij}\right|=0 \tag*{(2)}

in units where ℏ=c=1\hbar=c=1, which we assume henceforth. The dispersion relation has the general form ω2−k2=μ2(k,λ)\omega^2-k^2=\mu^2(\mathbf{k},\lambda) where k\mathbf{k} is the wave vector and, setting ωp=0\omega_p=0 for the sake of presentation simplicity, [7]

μ2(k,λ)=b22−λsgn⁡(k⋅b)(k⋅b)2+b2/4.(3)\mu^2(\mathbf{k},\lambda)=\frac{b^2}{2}-\lambda\operatorname{sgn}(\mathbf{k}\cdot\mathbf{b})\sqrt{(\mathbf{k}\cdot\mathbf{b})^2+b^2/4}. \tag*{(3)}

Here λ=±1\lambda=\pm1 is the photon circular polarization. Let β\beta be the angle between b\mathbf{b} and k\mathbf{k}, see Figure 2. There are two branches of the dispersion relation: (i) the gapped branch: λcos⁡β>0\lambda\cos\beta>0, μ2>0\mu^2>0 and (ii) the gapless branch λcos⁡β<0\lambda\cos\beta<0, μ2<0\mu^2<0. These are depicted in Figure 1. Photon radiation is associated with the latter.

Figure 1. Chiral branches of the photon dispersion relation. Arbitrary units.

The geometry of the photon radiation

Figure 2. The geometry of the photon radiation. The three vectors are generally not in the same plane, however the azimuthal angle is not shown. At high energy ϑ≪β\vartheta\ll\beta, i.e. β\beta is approximately the angle between pp and bb (angle of incidence).

Photon radiation.— The initial and final states of electron are solutions to the Dirac equation with energies EE, E′E' and momenta p\mathbf{p}, p′\mathbf{p}'. The wave function of the radiated photon with energy and momentum ω\omega, k\mathbf{k} and polarization λ\lambda reads [8]

A=e(kv2ω2ϵijei∗ej∗)1/2eik⋅x−iωt(4)\mathbf{A}=\mathbf{e}\left(\frac{k v}{2\omega^2\epsilon_{ij}e_i^*e_j^*}\right)^{1/2}e^{i\mathbf{k}\cdot\mathbf{x}-i\omega t} \tag*{(4)}

where e\mathbf{e} is the polarization vector for a given λ\lambda and v=∂ω/∂kv=\partial\omega/\partial k is the group velocity. The explicit expression for e\mathbf{e} can be found in [3]. Computing the transition amplitude e−→e−γe^-\to e^-\gamma and using the standard procedure to obtain the transition rate we obtain for the number of photons emitted per unit time in unit solid angle in the frequency interval dωd\omega [3]:

dWdΩdω=α16π∑δ(ω+E′−E)k3EE′ω2εijej∗ej[4ei∗ej(pipj′+pjpi′+δij(EE′−p⋅p′−m2))](5)\frac{dW}{d\Omega d\omega}=\frac{\alpha}{16\pi}\sum\delta(\omega+E'-E)\frac{k^3}{EE'\omega^2\varepsilon_{ij}e_j^*e_j} \left[4e_i^*e_j\left(p_i p_j'+p_j p_i'+\delta_{ij}(EE'-p\cdot p'-m^2)\right)\right] \tag*{(5)}

where p′=p−kp'=p-k. The energy conservation expressed by the delta-function is satisfied only for the gapless branch μ2<0\mu^2<0. Similarly to the conventional Cherenkov radiation, the Chiral Cherenkov one is radiated even by a particle moving with constant velocity, which corresponds to the limit ω≪E\omega\ll E when electron’s recoil is small. In this case, the radiating electron moves in phase with the wave. As the result both types of Cherenkov radiation have infinite coherence length in contrast to the bremsstrahlung.

Thus far we discussed the Chiral Cherenkov radiation by a single electron. A typical electric current pulse consists of N=1011N=10^{11} electrons (equivalent to the electron bunch charge 16 nC). The corresponding power

dI=N dP.dI=N\,dP.

The explicit form of the rate is rather bulky. We therefore consider two special cases.

High energy approximation.— Assume that E≫mE\gg m and ω≫∣μ∣\omega\gg|\mu|. The latter condition is equivalent to ω≫b,ωp\omega\gg b,\omega_p. The Chiral Cherenkov radiation rate (5) can then be written in a compact form [1, 2, 3]:

dWdωdΩ=αω2πEδ(ω2ϑ2+κ){fE(ω22E2−ωE+1)−m2ωE}θ(−κ),(6)\frac{dW}{d\omega d\Omega}=\frac{\alpha\omega}{2\pi E}\delta(\omega^2\vartheta^2+\kappa)\left\{fE\left(\frac{\omega^2}{2E^2}-\frac{\omega}{E}+1\right)-m^2\frac{\omega}{E}\right\}\theta(-\kappa), \tag*{(6)}

where f=λbcos⁡βf=\lambda b\cos\beta, and β\beta is now the angle between the incident electron direction and the vector bb since in this approximation the angle ϑ\vartheta between kk and pp is small, see Figure 2.

In the right-hand-side θ\theta is the step function and we defined

κ=−(1−ωE)ωf+m2ω2E2.(7)\kappa=-\left(1-\frac{\omega}{E}\right)\omega f+\frac{m^2\omega^2}{E^2}. \tag*{(7)}

The kinematic condition κ<0\kappa<0 can be satisfied only if f>0f>0. Since μ2≈−fω\mu^2\approx-f\omega it is equivalent to μ2<0\mu^2<0. Additionally, it imposes the upper limit on the photon spectrum:

ω<E(1+m2fE)−1.(8)\omega<E\left(1+\frac{m^2}{fE}\right)^{-1}. \tag*{(8)}

The lower limit is proportional to ωp\omega_p, but in the high-energy approximation it is neglected. The differential spectrum of the radiation power is given by

dPdωdΩ=ωdWdωdΩ.(9)\frac{dP}{d\omega d\Omega}=\omega\frac{dW}{d\omega d\Omega}. \tag*{(9)}

solid lines in Figure 3,Figure 4,Figure 5 display the high-energy approximation of the radiated photon spectrum, its angular distribution and dependence on angle of incidence (viz. angle between the electron momentum pp and the vector bb).

Figure 3. Intensity spectrum of Chiral Cherenkov radiation by a pulse of N=1011N=10^{11} electrons of energy E=3E=3 MeV, β=0\beta=0 in a Weyl semimetal with b=3.8b=3.8 eV (green), b=0.19b=0.19 eV (orange), and b=0.01b=0.01 eV (blue). Solid line: Eq. (6), dashed line: Eq. (14) with ωp=0.6\omega_p=0.6 meV, dotted line: bremsstrahlung spectrum in Co3_3Sn2_2S2_2 computed according to [9]. 1/cm ≈\approx 0.03 THz.

Figure 4. Angular distribution of Chiral Cherenkov radiation. N=1011N = 10^{11}, E=3E = 3 MeV, β=0\beta= 0, b=3.8b = 3.8 eV. Eq. (6) and Eq. (14) are indistinguishable.

Figure 5. Variation of Chiral Cherenkov radiation intensity with electron angle of incidence (the polar angle). The two loops have opposite polarization. N=1011N = 10^{11}, E=3E = 3 MeV, b=3.8b = 3.8 eV.

The photon emission rate Eq. (6) can be further simplified by noting that in semimetals m2≫fEm^2\gg fE:

dWdωdΩ≈αf2πωδ(ϑ2−fω+m2E2)θ(E2fm2−ω),(10)\frac{dW}{d\omega d\Omega}\approx\frac{\alpha f}{2\pi\omega}\delta\left(\vartheta^2-\frac{f}{\omega}+\frac{m^2}{E^2}\right)\theta\left(\frac{E^2f}{m^2}-\omega\right), \tag*{(10)}

where the step-function follows from (8). Multiplying by ω\omega and integrating over dΩ=πdϑ2d\Omega=\pi d\vartheta^2 we obtain the power spectrum

dPdω=αf2θ(E2fm2−ω),(11)\frac{dP}{d\omega}=\frac{\alpha f}{2}\theta\left(\frac{E^2f}{m^2}-\omega\right), \tag*{(11)}

and the the total power

P=αf2E22m2=αb2cos⁡2βE22m2.(12)P=\frac{\alpha f^2E^2}{2m^2}=\frac{\alpha b^2\cos^2\beta E^2}{2m^2}. \tag*{(12)}

Integrating (10) over ω\omega furnishes the angular distribution of the radiation

dWdΩ=αf2π1ϑ2+m2/E2.(13)\frac{dW}{d\Omega} = \frac{\alpha f}{2\pi}\frac{1}{\vartheta^2 + m^2/E^2}. \tag*{(13)}

Evidently, the total radiation power increases as b2b^2, while the UV cutoff of the spectrum as bb. It is therefore advantageous to have as high bb as possible to obtain the highest power of the Chiral Cherenkov radiation. It also guarantees, as per Figure 3,Figure 4, that the radiation power of the polarized photons by means of the Chiral Cherenkov mechanism is larger than that of the non-polarized photons through the conventional sources.

Electron incident along bb. — According to (12) the radiation power vanishes when the incident electron moves perpendicularly to the vector bb. Therefore, in practice one would like to position the Weyl semimetal sample such that p×b=0\mathbf{p} \times\mathbf{b} = 0. In this case (5) significantly simplifies. Performing the angular integral yields the rate:

dWdω=(ω2−μ2)(EE′−m2)−(2ωE−μ2)(2ωE′+μ2)4(ω2−μ2)ω2E∣p∣×α2(1+ωp2ω2)θ(ω+−ω),(14)\frac{dW}{d\omega} = \frac{(\omega^2-\mu^2)(EE'-m^2)-(2\omega E-\mu^2)(2\omega E'+\mu^2)}{4(\omega^2-\mu^2)\omega^2 E|\mathbf{p}|}\times\frac{\alpha}{2}\left(1+\frac{\omega_p^2}{\omega^2}\right)\theta(\omega_+-\omega), \tag*{(14)}

where μ2=ωp2−λbω\mu^2 = \omega_p^2-\lambda b\omega, p2=E2−m2\mathbf{p}^2 = E^2-m^2 and

ω+=[λb(2p2+ωp2)+2Eωp2+2∣p∣(λbE−ωp2)2−m2(b2+4ωp2)]×[λb(4E+λb)+4m2]−1.(15)\omega_+ = \left[\lambda b(2\mathbf{p}^2+\omega_p^2)+2E\omega_p^2+2|\mathbf{p}|\sqrt{(\lambda bE-\omega_p^2)^2-m^2(b^2+4\omega_p^2)}\right]\times\left[\lambda b(4E+\lambda b)+4m^2\right]^{-1}. \tag*{(15)}

In the high energy limit ω≫b\omega\gg b and E≫mE \gg m these equations reduce to those displayed in the previous section. Dashed lines in Figure 3,Figure 4 represent a calculation of intensity with Eq. (14). Evidently, the high-energy approximation is quite accurate all the way down to the IR region ω∼ωp\omega\sim\omega_p where our approximation for the dielectric constant ε\varepsilon breaks down anyway.

Dirac semimetals. —In Dirac semimetals b=0b = 0. However, by applying external parallel magnetic BB and electric E\mathcal{E} fields one can generate the chiral charge. The corresponding axial chemical potential reads [10]:

μ5=34π2v3e2EBT2τV,(16)\mu_5 = \frac{3}{4\pi^2}v^3 e^2\frac{\mathcal{E}B}{T^2}\tau_V, \tag*{(16)}

where τV\tau_V is the chirality relaxation time, and vv is the Fermi velocity of Dirac quasiparticles. The chiral anomaly manifests itself as the Chiral Magnetic Effect which is induction of electric current in the direction of the applied magnetic field:

jCM=σχB,(17)j_{\mathrm{CM}} = \sigma_\chi B, \tag*{(17)}

where the chiral magnetic conductivity is σχ=(2/π)αμ5\sigma_\chi= (2/\pi)\alpha\mu_5. In the presence of the current (17) the photon dispersion relation acquires a branch with μ2<0\mu^2 < 0 in a way very similar to Weyl semimetals. In fact all formulas derived in the high-energy limit, apply at finite μ5\mu_5 as well upon the substitution f=λσχf = \lambda\sigma_\chi [1, 2].

In practice, however, the values of σχ\sigma_\chi are much smaller than bb as can be seen in Figure 6. The separation of the Weyl nodes in a chiral semimetal is of the order $\Delta \sim 10−2(π/a)10^{-2}(\pi/a). For example, in TaAs it is Δ=80\Delta=80 eV [11, 12]. The corresponding bb-parameter in (1) is of magnitude b=0.19b=0.19 eV. In contrast, in a magnetic Weyl semimetal Co3Sn2S2\mathrm{Co}_3\mathrm{Sn}_2\mathrm{S}_2 b=3.8b=3.8 eV [5, 6].

Figure 6. The temperature and magnetic field dependence of μ5\mu_5 at E=1 mV/m\mathcal{E}=1\,\mathrm{mV}/\mathrm{m}. The chiral magnetic conductivity is σχ=4.6⋅10−3μ5\sigma_\chi=4.6\cdot10^{-3}\mu_5.

Summary.— The Chiral Cherenkov radiation appears due to anomalous response of chiral media to the electromagnetic field of a fast charge. In magnetic Weyl semimetals, for an intense electron beam of a few MeV energy, the intensity of the Chiral Cherenkov radiation can exceed the intensity of usual bremsstrahlung in the THz frequency range. Moreover, the THz CC radiation is circularly polarized, which can be very useful for practical applications.

Acknowledgements

We thank Sebastian Grieninger, Fangcheng He and Andrea Palermo for a useful discussion. This work was supported by the U.S. Department of Energy, Office of Science, National Quantum Information Science Research Centers, Co-design Center for Quantum Advantage (C2QA) under Contract No. DE-SC0012704 (KI, DK), and the U.S. Department of Energy under Grants DE-FG88ER41450 (DK) and DE-SC0012704 (DK), This work of JH and KT was supported in part by the U.S. Department of Energy Grant No. DE-SC0023692. The work of QL was supported by the U.S. Department of Energy, Office of Basic Energy Sciences, Division of Materials Sciences and Engineering, under contract No. DE-SC0012704.

References

References

  1. [1]K. Tuchin, “Radiative instability of quantum electrodynamics in chiral matter,” Phys. Lett. B 786 (2018) 249–254, arXiv:1806.07340 [hep-ph].DOI
  2. [2]J. Hansen and K. Tuchin, “Collisional energy loss and the chiral magnetic effect,” Phys. Rev. C 104 no. 3, (2021) 034903, arXiv:2012.06089 [hep-ph].DOI
  3. [3]K. Tuchin, “Chiral Cherenkov and chiral transition radiation in anisotropic matter,” Phys. Rev. D 98 no. 11, (2018) 114026, arXiv:1809.08181 [hep-ph].DOI
  4. [4]X.-G. Huang and K. Tuchin, “Transition Radiation as a Probe of the Chiral Anomaly,” Phys. Rev. Lett. 121 no. 18, (2018) 182301, arXiv:1808.00635 [hep-ph].DOI
  5. [5]M. A. Zubkov, “Weyl orbits as probe of the chiral magnetic separation effect in magnetic Weyl semimetals,” arXiv:2311.12712 [cond–mat.mes–hall].
  6. [6]Q. Wang, Y. Xu, R. Lou, Z. Liu, M. Li, Y. Huang, D. Shen, H. Weng, S. Wang, and H. Lei, “Large intrinsic anomalous hall effect in half-metallic ferromagnet co3sn2s2 with magnetic weyl fermions,” Nature communications 9 no. 1, (2018) 1–8.
  7. [7]Z. Qiu, G. Cao, and X.-G. Huang, “On electrodynamics of chiral matter,” Phys. Rev. D 95 no. 3, (2017) 036002, arXiv:1612.06364 [cond–mat.mes–hall].DOI
  8. [8]A. Alekseev and Y. P. Nikitin, “Quantization of the electromagnetic field in a dispersive medium,” Sov. Phys. JETP 23 (1966) 608.
  9. [9]Particle Data Group Collaboration, R. L. Workman and Others, “Review of Particle Physics,” PTEP 2022 (2022) 083C01.
  10. [10]K. Fukushima, D. E. Kharzeev, and H. J. Warringa, “The Chiral Magnetic Effect,” Phys. Rev. D 78 (2008) 074033, arXiv:0808.3382 [hep-ph].DOI
  11. [11]S. Y. Xu et al., “Discovery of a Weyl Fermion semimetal and topological Fermi arcs,” Science 349 (2015) 613–617, arXiv:1502.03807 [cond–mat.mes–hall].DOI
  12. [12]B. Q. Lv et al., “Experimental discovery of Weyl semimetal TaAs,” Phys. Rev. X 5 no. 3, (2015) 031013, arXiv:1502.04684 [cond–mat.mtrl–sci].DOI

Paper details

vlnc:W_m7gpq3mt
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents