betaFeedback

WorksPhysics and Astronomy

Impedance-controlled Laser Dynamics in Near-Zero Index Media

Abstract

Near-zero-index (NZI) media have been explored for controlling spontaneous emission, but their impact on stimulated emission and lasing properties remains unexplored within their different classes, i.e., epsilon-near-zero (ENZ), mu-near-zero (MNZ), or epsilon-and-mu-near-zero (EMNZ). Here, we show theoretically that lasing emission depends on the class of NZI medium and can be controlled through its impedance. The major difference, as compared to conventional laser media, arises from the incorporation of dispersive electric and magnetic properties in the derivation of the Maxwell Boch equations. In this framework, the analysis of laser rate equation reveals that the ENZ class, with high impedance, can reduce the lasing threshold through enhanced electric-field interaction. Numerical simulations further show that the NZI class also governs the spectral response: whereas ENZ and MNZ classes suffer from multimode emission, the EMNZ class preserves single-mode emission.

The electromagnetic environment plays a central role in shaping light–matter interactions. Since the pioneering work of Purcell [1], it is known that the spontaneous emission rate of an emitter can be strongly modified by its environment, either through optical confinement in cavities [2] or through the engineering of the local electromagnetic density of states in structured media [3].

In this context, near-zero-index (NZI) media have been shown to strongly modify spontaneous emission in a manner that depends on the specific NZI class and dimensionality [4, 5, 6, 7, 8, 9, 10, 11]. NZI media are classified as epsilon-near-zero (ENZ), mu-near-zero (MNZ), or epsilon-and-mu-near-zero (EMNZ), depending on whether the relative permittivity, permeability, or both approach zero [12, 13, 14, 15, 16]. This classification gives rise to markedly different electromagnetic responses: ENZ and MNZ media can support slow-light propagation, while their impedances tend toward infinity and zero, respectively, for infinitely homogeneous media [4, 17, 18]. In contrast, both the impedance and group refractive index of EMNZ media can be tuned through their dispersive properties [19, 20, 21, 22, 23]. These differences strongly affect the balance between electric- and magnetic-field interactions, leading to their pronounced enhancement in ENZ and MNZ media, respectively [14, 24]. Such unusual electromagnetic responses have been extensively exploited to control spontaneous emission, enabling enhanced emission rates [4, 5, 6, 7, 8, 9] and long-range collective effects such as superradiance [25, 26, 27, 28, 29].

Despite these advantages, the effects of NZI environments on stimulated emission and laser dynamics remain largely unexplored [4, 21]. For instance, it remains unclear how the individual NZI classes influence the lasing threshold, modal properties, and laser dynamics. Addressing this question requires disentangling the respective roles of the dispersive permittivity, permeability, group velocity, and wave impedance in light–matter interactions.

In this Letter, we investigate theoretically how the class of NZI media governs the threshold and spectral properties of solid-state lasers. By incorporating the electric and magnetic dispersive properties of the active medium analytically into the laser rate equations, we reveal that light–matter interactions are governed not only by the group refractive index, but also by the wave impedance. Specifically, the high impedance of ENZ media, combined with their ability to slow down light velocity, enhances the electric-field interactions and reduces the lasing threshold, whereas the low impedance of MNZ media weakens the electric-field interactions and increases the lasing threshold. Likewise, the lasing threshold in EMNZ media can be controlled through their impedance. Using finite-difference time-domain (FDTD) simulations, we further examine the spectral response. The large group indices of ENZ and MNZ media lead to multimode emission, whereas EMNZ media preserve single-mode lasing despite their near-zero refractive index. Together, these results reveal how the electromagnetic parameters of NZI media, i.e. group index and impedance, independently shape the threshold and spectral properties of lasers.

To disentangle the respective roles of the active medium permittivity, permeability, and group index in laser emission, we generalized the Maxwell–Bloch formalism from which laser rate equations are commonly derived. The active medium is modeled as an ensemble of two-level emitters described by conventional Bloch equations for the population inversion and microscopic polarization [30, 31]. However, the frequency-dependent permittivity and permeability of the active medium are explicitly retained in Maxwell’s equations, thereby modifying the effective coupling between the electromagnetic field and the atomic polarization.

The electromagnetic field is considered inside an optical linear cavity filled with this dispersive active medium. Expanding the electromagnetic response around the laser frequency and applying the slowly varying envelope approximation [30, 32] (see Appendix § 2), the temporal evolution of the electric field amplitude E(t)E(t) of the resonant cavity mode [32] is described by

dE(t)dt=ivg(kp,0−kp,c)E(t)−γc2E(t)+iμvgω02ε0cPat(t)(1)\begin{aligned} \frac{dE(t)}{dt} &= iv_g(k_{p,0}-k_{p,c})E(t) \\ &\quad- \frac{\gamma_c}{2}E(t) + i\sqrt{\frac{\mu v_g\omega_0}{2\varepsilon_0 c}}P_{at}(t) \tag*{(1)} \end{aligned}

where vgv_g is the group velocity, kp,0=ω0n(ω0)/ck_{p,0}=\omega_0 n(\omega_0)/c and kp,c=ωcn(ωc)/ck_{p,c}=\omega_c n(\omega_c)/c are the wave vectors related to the propagation ω0\omega_0 and cavity resonance ωc\omega_c angular frequencies, respectively, nn is the phase refractive index, γc\gamma_c is the cavity loss rate, and PatP_{at} is the atomic polarization. The first term in Eq. (1) describes the frequency detuning between the atomic transition and the cavity resonance. In conventional media, one typically assumes ng≃nn_g \simeq n, such that vg≃c/nv_g \simeq c/n and this first term reduces to the usual form i(ω0−ωc)E(t)i(\omega_0-\omega_c)E(t). This approximation, however, is no longer valid in NZI media, where the group index ngn_g can differ drastically from the phase index nn and the detuning must therefore retain its explicit dependence on the group velocity. In conventional laser rate equations, the light–matter coupling, i.e., the third term of Eq. (1), is described by the coefficient vgiω0/2nε0cv_g i\omega_0/2n\varepsilon_0c, where the phase refractive index nn appears in place of the wave impedance Z=μ/εZ=\sqrt{\mu/\varepsilon} [30, 32]. By explicitly accounting for the frequency-dependent magnetic permeability, the wave impedance emerges as the parameter governing the strength of the light–matter coupling.

From these dispersive Maxwell–Bloch equations, the laser rate equations can be derived following the usual procedure for a semi-conductor laser hereafter considered [30] (see Appendix § 3). The resulting carrier NN and photon SS density dynamics are given by

dNdt=Λp−Zvgg(N)S−γnrN,dSdt=nμβγspN+Zvgg(N)S−γcS,(2)\begin{aligned} \frac{dN}{dt} &= \Lambda_p-Zv_gg(N)S-\gamma_{nr}N,\\ \frac{dS}{dt} &= n\mu\beta\gamma_{sp}N+Zv_gg(N)S-\gamma_cS, \tag*{(2)} \end{aligned}

where Λp\Lambda_p is the pumping rate, g(N)=g0(N−N0)g(N)=g_0(N-N_0) is the material gain (in m−1\mathrm{m}^{-1}) with g0g_0 the gain constant (in m2\mathrm{m}^2) and N0N_0 the carrier density at transparency, γnr\gamma_{nr} is the nonradiative recombination rate, γsp\gamma_{sp} and β\beta are the spontaneous decay rate and beta factor (i.e., the fraction of spontaneous emission coupled to the lasing cavity mode), respectively, and γc\gamma_c is the cavity loss rate. The ε\varepsilon- and μ\mu-dependence of γsp\gamma_{sp} is taken from the dispersive derivations of Ref. [4]. The corresponding dependence of stimulated emission derived herein depends explicitly on ng=c/vgn_g=c/v_g and ZZ, in agreement with previously reported descriptions of stimulated Einstein coefficient in dispersive media, i.e., B21′=1ngμεB21B'_{21}=\frac{1}{n_g}\sqrt{\frac{\mu}{\varepsilon}}B_{21} [4, 33]. The above formalism therefore provides a framework to disentangle the respective contributions of dispersive permittivity, permeability, group velocity and wave impedance to laser emission.

We first solve the laser rate equations, Eqs. (2), for a conventional InGaAsP active medium, used as reference. The material parameters are fixed to g0=2.50×10−16cm2g_0=2.50\times10^{-16}\mathrm{cm}^2, N0=1.00×1018cm−3N_0=1.00\times10^{18}\mathrm{cm}^{-3}, γnr−1=γsp−1=2.20ns\gamma_{nr}^{-1}=\gamma_{sp}^{-1}=2.20\mathrm{ns}, β=1.00×10−4\beta=1.00\times10^{-4}, and γc−1=1.60ps\gamma_c^{-1}=1.60\mathrm{ps} [37], with the active medium modeled as a dielectric with n=3.40n=3.40 and ng=4.00n_g=4.00 [34]. We then compare this reference with hypothetical dispersive media described by a Drude-Lorentz model, while keeping all other rate-equation parameters unchanged [14]. The parameters of ε\varepsilon and μ\mu models are chosen to match the target values at the lasing frequency, with the group index calculated from their frequency derivatives. We first consider two limiting cases: an ENZ medium with ε=0.01\varepsilon=0.01 and μ=1.00\mu=1.00, and an MNZ medium with ε=1.00\varepsilon=1.00 and μ=0.01\mu=0.01. Although both media share the same large group index, (ng=22.01n_g=22.01), their impedances differ substantially (ZENZ=10.00Z_{ENZ}=10.00 and ZMNZ=0.10Z_{MNZ}=0.10), enabling us to isolate the role of impedance in cavity dynamics under slow-light conditions (see Table I).

Figure 1(a) shows the steady-state laser output power as a function of injection current. The ENZ medium reduces the lasing threshold by approximately a factor of two (Ith=7.52mAI_{th}=7.52\mathrm{mA}) relative to the reference (Ith=15.52mAI_{th}=15.52\mathrm{mA}), whereas the MNZ medium doubles it (Ith=31.52mAI_{th}=31.52\mathrm{mA}). Despite their identical group indices, ENZ and MNZ therefore produce opposite shifts of the lasing threshold, indicating that slow light with the same group velocity can interact differently with matter depending on the impedance. The high impedance of the ENZ medium enhances the electric field, thereby strengthening the electric-dipole interactions with the active medium, whereas the low impedance of the MNZ medium enhances the magnetic field and thus favors the magnetic response at the expense of the electric-dipole interactions.

Light–current characteristics for reference InGaAsP, ENZ, MNZ, and EMNZ media

Figure 1. Light–current characteristics obtained from solving the laser rate equations for (a) the reference InGaAsP (blue), ENZ (red), and MNZ (yellow) media with a schematic of the laser linear cavity as inset, and (b) the reference InGaAsP and two EMNZ media (green and light blue). Dashed lines indicate the corresponding lasing thresholds.

The role of the impedance is further highlighted by the EMNZ class, for which the impedance depends on the relative dispersion of both ε\varepsilon and μ\mu [4]. We therefore consider two EMNZ media with different combinations of permittivity and permeability, EMNZ-1 with ε=0.01\varepsilon= 0.01, μ=0.10\mu= 0.10 (ZEMNZ−1=3.16Z_{\mathrm{EMNZ-1}} = 3.16) and EMNZ-2 with ε=0.20\varepsilon= 0.20, μ=0.01\mu= 0.01 (ZEMNZ−2=0.22Z_{\mathrm{EMNZ-2}} = 0.22). As shown in Figure 1(b), the high-impedance configuration, i.e. EMNZ-1 case, which exhibits the lower group index, lowers the lasing threshold, whereas the low-impedance configuration, i.e. EMNZ-2 case, increases it (see Table I). This demonstrates that a more slowed down light does not necessarily enhance light–matter interactions, consistently with the above-mentioned impedance control of light–matter interaction, i.e. the ENZ/MNZ cases.

The threshold trends observed in Figure 1 are further confirmed by studying the carrier-density and photon density dynamics as shown in Figure 2. Since the threshold injection current scales with the threshold carrier density NthN_{\mathrm{th}}, the latter provides a direct measure of the threshold behavior. From Eqs. (2), the threshold carrier density is given by

Image not extracted

Figure 2. Temporal evolutions of (a-b) the carrier density (NN) normalized by the transparency carrier density (N0N_0) and (c-d) the intracavity photon density (SS) for an injection current of 100 mA. The blue, red, and yellow curves correspond to the reference InGaAsP medium, the ENZ medium, and the MNZ medium, respectively, while the light-blue and green curves correspond to the two EMNZ media. Dashed lines indicate the corresponding thresholds for MNZ.

Nth=N0+1Zgthg0,(3)N_{\mathrm{th}} = N_0 + \frac{1}{Z}\frac{g_{\mathrm{th}}}{g_0}, \tag*{(3)}

where gthg_{\mathrm{th}} is the threshold gain. This relation shows that, within ε\varepsilon- and μ\mu-dispersive media, the threshold carrier density scales as the inverse of the wave impedance ZZ (see Appendix § 4 for the steady-state analysis of Eqs (2)).

RefENZMNZEMNZ-1EMNZ-2
ε\varepsilon11.560.011.000.010.20
μ\mu1.001.000.010.100.01
ngn_g4.0022.0122.017.6110.26
ZZ0.2910.000.103.160.22
IthI_{\mathrm{th}} (mA)15.527.5231.528.0118.12
Nth/N0N_{\mathrm{th}}/N_02.131.034.331.102.49
SthS_{\mathrm{th}}0.835.013.711.722.06
(1016 cm−3)(10^{16}\,\mathrm{cm}^{-3})

Table 1. Summary of the electromagnetic parameters and the corresponding lasing IthI_{\mathrm{th}}, carrier density NthN_{\mathrm{th}}, and photon density SthS_{\mathrm{th}} thresholds for the different configurations considered.

Figure 2(a) shows the temporal evolution of the carrier density for the reference, ENZ, and MNZ media. The ENZ medium reaches the lasing regime at a carrier density close to N0N_0, with laser relaxation oscillations rapidly stabilizing, whereas the MNZ medium exhibits much larger carrier-density oscillations around a higher threshold. This contrasting dynamic reflects the different light–matter interactions in the two NZI classes: in ENZ medium, the enhanced electric field strengthens the interaction with the atomic polarization, allowing lasing to develop with only small variations of the population inversion. In MNZ medium, the weaker electric-field interactions require larger variations of the population inversion to produce a polarization response that is strong enough for efficient coupling to the optical field. The same behavior applies to the EMNZ class, where the carrier-density dynamics likewise depend on the impedance of the medium (Figure 2(b)).

The field enhancement effect is further evidenced by the temporal evolution of the photon density SS shown in Figure 2(c,Figure 2d). The larger group index of the ENZ and MNZ media leads to an increased photon accumulation in the cavity, resulting in a higher threshold photon density SthS_{\mathrm{th}} (Figure 2(c)) (see Appendix § 4). The asymmetry between these two slow-light regimes arises from the impedance-dictated energy storage channel: the electromagnetic energy is stored predominantly in the electric field for ENZ and in the magnetic field for MNZ. For the EMNZ media (Figure 2(d)), the group index is likewise larger than the reference medium, leading to an increased SthS_{\mathrm{th}} in both cases, while the dominant energy-storage channel depends on the wave impedance. Overall, these results show that the NZI class, through its combined influence on group index and impedance, changes the carrier and photon dynamics and, consequently, the lasing threshold.

To investigate the impact of the NZI class on the laser emission spectrum, we performed full-wave 2D-FDTD simulations with transverse-electric polarization [35]. The active medium is described by a dispersive electromagnetic response, in which both the permittivity and permeability follow frequency-dependent constitutive relations [19, 35, 36, 37], with a self-consistent four-level gain model coupled to Maxwell's equations [35, 38, 39]. The laser cavity is modeled as a one-dimensional Fabry–Perot cavity, with the active medium enclosed between two Bragg mirrors composed of alternating InGaAsP (n=3.40n = 3.40)

and air layers satisfying the quarter-wave condition. The target wavelength λ0=1.3 μm\lambda_0 = 1.3\,\mu\mathrm{m} is located at the center of the mirror’s photonic band gap, and the whole cavity is embedded in air and terminated by perfectly matched layers.

The emission spectra (Figure 3) reveal distinct modal behaviors across the NZI classes. The reference InGaAsP cavity exhibits a single emission peak near 1.3μm, corresponding to stable single-mode operation. In contrast, the ENZ and MNZ media exhibit additional longitudinal modes around the lasing frequency, characteristics of multimode emission. The EMNZ cavity retains a dominant single emission peak. This behavior is governed by the cavity mode spacing, Δν=c/(2ngL)\Delta\nu= c/(2n_gL) [30]: the large group index of ENZ and MNZ (ng=22.01n_g = 22.01) reduces the longitudinal mode spacing, allowing additional modes to emerge, whereas the moderate group index of the EMNZ preserves a sufficiently large mode spacing for single-mode operation to prevail. The frequency shift observed in the EMNZ-2 medium with the lowest impedance arises from the dispersion of ϵ\epsilon and μ\mu. The peak occurs where the permittivity approaches zero, indicating that, in the cavity mode competition, the mode with the lowest permittivity, and hence the highest impedance, is preferentially favored.

Emission spectra around 1.3 micrometres

Figure 3. Emission spectra around 1.3μm (dashed line) obtained from FDTD simulations for an injection rate of 1016 s−110^{16}\,\mathrm{s}^{-1}. The blue, red, yellow, green and light blue curves correspond to the reference InGaAsP, ENZ, MNZ, EMNZ-1 and EMNZ-2 media, respectively. The cavity lengths are L=5λ0/nL = 5\lambda_0/n for the reference and L=λ0/(2n)L = \lambda_0/(2n) for the ENZ, MNZ and EMNZ media.

Interestingly, the single-mode emission observed for the EMNZ media and the multimode emission of the ENZ/MNZ media are consistent with previously reported experiments. Contractor et al. demonstrated that a photonic-crystal laser featuring a Dirac cone at the Γ\Gamma-point maintained single-mode operation over large cavity areas, whereas lifting the Dirac-cone degeneracy led to multimode emission [40]. Such Dirac-cone photonic crystals have been shown to exhibit an effective EMNZ response, while lifting the degeneracy drives the system toward ENZ or MNZ responses [12, 41, 42]. Our results provide a unified interpretation of these observations through the group index dependence of the NZI classes.

In conclusion, we have theoretically demonstrated that each of the three classes of NZI media provides its own direct control over laser emission properties. By developing laser rate equations that explicitly incorporate the electric and magnetic dispersive properties of the active medium, we revealed that the impedance is the key parameter governing the lasing threshold. The high impedance of ENZ media lowers the lasing threshold through enhanced electric-field interactions, whereas the low impedance of MNZ media makes them less favorable for low-threshold lasing due to their weaker electric-field-mediated light–matter interactions. In EMNZ media, the tunable impedance provides an additional degree of control over the lasing threshold. Numerical simulations further showed that the slow-light response of ENZ and MNZ media promotes the emergence of additional spectral modes due to a reduced cavity-mode spacing, whereas the lower group index of EMNZ media helps preserve single-mode emission. These results establish a direct link between the electromagnetic properties of NZI media and laser operation, opening new opportunities for the design of photonic-crystal lasers [40], where Dirac-cone dispersions, for instance, provide a natural route toward effective NZI responses [42]. Furthermore, the extension of this approach to time-varying metamaterials could enable dynamic control of lasing properties through actively tunable electromagnetic responses [43, 44, 45].

Acknowledgments

The authors would like to thank J. Faist for stimulating discussion. A. D. is a research fellow of the Fonds de la Recherche Scientifique – FNRS. M.L. is a Research Associate of the Fonds de la Recherche Scientifique – FNRS. This research used resources of the "Plateforme Technologique de Calcul Intensif (PTCI)" (http://www.ptci.unamur.be) located at the University of Namur, Belgium, which is supported by the FNRS-FRFC, the Walloon Region, and the University of Namur (Conventions No. 2.5020.11, GEQ U.G006.15, 1610468, RW/GEQ2016 et U.G011.22). The PTCI is member of the "Consortium des Équipements de Calcul Intensif (CÉCI)" (http://www.ceci-hpc.be). Generative AI tools were used during the preparation of this manuscript to assist with text drafting, language editing, and for translation purposes only. All AI-generated content was subsequently reviewed, modified, and validated by the authors, who take full responsibility for the final content of the manuscript.

Maxwell wave equation in a dispersive cavity

We derive hereafter the evolution equation of the electric field E\boldsymbol{\mathcal{E}} inside a laser cavity containing a dispersive medium. We start from the wave equation in a dispersive medium [46],

∇2E(r,t)−n2(ω)c2∂2E(r,t)∂t2=μ(ω)ε0c2∂2Pat(r,t)∂t2,(4)\nabla^2\boldsymbol{\mathcal{E}}(\boldsymbol{r},t)-\frac{n^2(\omega)}{c^2}\frac{\partial^2\boldsymbol{\mathcal{E}}(\boldsymbol{r},t)}{\partial t^2}=\frac{\mu(\omega)}{\varepsilon_0c^2}\frac{\partial^2\boldsymbol{P}_{\mathrm{at}}(\boldsymbol{r},t)}{\partial t^2}, \tag*{(4)}

where ω\omega is the angular frequency, cc the speed of light and ε0\varepsilon_0 the vacuum permittivity. The phase index n(ω)n(\omega) and permeability μ(ω)\mu(\omega) account for the dispersive properties of the medium and Pat\boldsymbol{P}_{\mathrm{at}} is the atomic polarization. Inside the cavity, the electric field is expanded over the cavity eigenmodes as

E(r,t)=e−iω0t∑qEq(t)eikp,q⋅ruq(r),(5)\boldsymbol{\mathcal{E}}(\boldsymbol{r},t)=e^{-i\omega_0t}\sum_q E_q(t)e^{i\boldsymbol{k}_{p,q}\cdot\boldsymbol{r}}\boldsymbol{u}_q(\boldsymbol{r}), \tag*{(5)}

where Eq(t)E_q(t), kp,q\boldsymbol{k}_{p,q} and uq(r)\boldsymbol{u}_q(\boldsymbol{r}) denotes the slowly varying field envelope, the complex wavevector and the spatial distribution of mode qq, respectively [32]. Each cavity mode induces a corresponding polarization Pq(t)\boldsymbol{P}_q(t), allowing the atomic polarization Pat\boldsymbol{P}_{\mathrm{at}} to be expanded on the same modal basis [32]. Substituting these modal expansions into the wave equation and working in the frequency domain yields an evolution equation for the modal amplitudes Eq(ω)E_q(\omega).

Thanks to the slowly varying envelope approximation [30, 32], spatial derivatives of the slowly varying envelopes can be neglected with respect to those of the rapidly varying phase factors, such that

∇2[eikp,q⋅ruq(r)]≃−kp,q2eikp,q⋅ruq(r).(6)\nabla^2\left[e^{i\boldsymbol{k}_{p,q}\cdot\boldsymbol{r}}\boldsymbol{u}_q(\boldsymbol{r})\right]\simeq-k_{p,q}^2e^{i\boldsymbol{k}_{p,q}\cdot\boldsymbol{r}}\boldsymbol{u}_q(\boldsymbol{r}). \tag*{(6)}

The propagation wavevector kp\boldsymbol{k}_p is first expanded to second order around the central frequency ω0\omega_0, then squared and only the lowest order derivatives are retained:

kp2(ω)≃kp,02+2vgkp,0(ω−ω0)+12kp,0σ(ω−ω0)2(7)k_p^2(\omega) \simeq k_{p,0}^2+\frac{2}{v_g}k_{p,0}(\omega-\omega_0)+\frac{1}{2}k_{p,0}\sigma(\omega-\omega_0)^2 \tag*{(7)}

where kp,0=kp(ω0)k_{p,0}=k_p(\omega_0), vg=dkpdω∣ω0v_g=\left.\frac{d k_p}{d\omega}\right|_{\omega_0} is the group velocity and σ=d2kpdω2∣ω0\sigma=\left.\frac{d^2 k_p}{d\omega^2}\right|_{\omega_0} is the group-velocity-dispersion coefficient. Assuming that the polarization envelope varies also slowly in time, the material response is evaluated at the central frequency by the approximate expression:

μ(ω)ω2Pq(ω−ω0)≃μ(ω0)ω02Pq(ω−ω0).(8)\mu(\omega)\omega^2P_q(\omega-\omega_0)\simeq\mu(\omega_0)\omega_0^2P_q(\omega-\omega_0). \tag*{(8)}

After performing inverse Fourier transformation, the modal amplitudes satisfy

∑q[(kp,02−kp,q2)Eq+2ikp,0vgdEqdt+kp,0σd2Eqdt2]eikp,q⋅ruq(r)=−μ(ω0)ω02ε0c2∑qPqeikp,q⋅ruq(r).(9)\sum_q\left[(k_{p,0}^2-k_{p,q}^2)E_q+\frac{2ik_{p,0}}{v_g}\frac{dE_q}{dt}+k_{p,0}\sigma\frac{d^2E_q}{dt^2}\right]e^{i\mathbf{k}_{p,q}\cdot\mathbf{r}}\mathbf{u}_q(\mathbf{r}) =-\frac{\mu(\omega_0)\omega_0^2}{\varepsilon_0c^2}\sum_qP_qe^{i\mathbf{k}_{p,q}\cdot\mathbf{r}}\mathbf{u}_q(\mathbf{r}). \tag*{(9)}

We now assume that the field frequency is close to a single cavity resonance and that mode spacing is sufficiently large to neglect all other modes [32]. The modal expansion therefore reduces to a single mode characterized by the complex wavevector k~p,c\tilde{k}_{p,c}. The field equation (9) becomes

1vgdEdt−i(kp,02−k~p,c2)2kp,0E=iμ(ω0)ω02ε0n(ω0)cPat.(10)\frac{1}{v_g}\frac{dE}{dt}-\frac{i(k_{p,0}^2-\tilde{k}_{p,c}^2)}{2k_{p,0}}E=\frac{i\mu(\omega_0)\omega_0}{2\varepsilon_0n(\omega_0)c}P_{at}. \tag*{(10)}

where we additionally consider a negligible group-velocity-dispersion coefficient (σ≃0\sigma\simeq0). Cavity losses are introduced through the imaginary part of the resonance wavevector, k~p,c=kp,c−iαc2\tilde{k}_{p,c}=k_{p,c}-\frac{i\alpha_c}{2}, where αc\alpha_c is the total cavity loss coefficient. Defining the cavity decay rate (γc=vgαc\gamma_c=v_g\alpha_c), we finally obtain

dEdt=ivg(kp,0−kp,c)E−γc2E+vgiμ(ω0)ω02ε0n(ω0)cPat.(11)\frac{dE}{dt}=iv_g(k_{p,0}-k_{p,c})E-\frac{\gamma_c}{2}E+v_g\frac{i\mu(\omega_0)\omega_0}{2\varepsilon_0n(\omega_0)c}P_{at}. \tag*{(11)}

which corresponds to (1) of the main text with Z(ω0)=μ(ω0)/n(ω0)Z(\omega_0)=\mu(\omega_0)/n(\omega_0).

From Maxwell-Bloch equations to laser rate equations

We briefly show how the semiconductor laser rate equations can be deduced from the Maxwell–Bloch equations. In a semiconductor, the microscopic quantities, i.e., electrons, possess de Broglie wavevector ke\mathbf{k}_e [30]. The microscopic polarization pkep_{\mathbf{k}_e}, the electronic transition frequency ωa,ke\omega_{a,\mathbf{k}_e}, the detuning δke=ω0−ωa,ke\delta_{\mathbf{k}_e}=\omega_0-\omega_{a,\mathbf{k}_e}, the dipole moment dked_{\mathbf{k}_e}, and the carrier populations in the conduction and valence bands therefore become dependent on ke\mathbf{k}_e. The population inversion wkew_{\mathbf{k}_e} is written as

wke=Ne,ke+Nh,ke−1,(12)w_{\mathbf{k}_e}=N_{e,\mathbf{k}_e}+N_{h,\mathbf{k}_e}-1, \tag*{(12)}

where Ne,keN_{e,\mathbf{k}_e} and Nh,keN_{h,\mathbf{k}_e} denote the occupation probabilities of electrons and holes, respectively. The semiconductor laser Bloch equations then read [30]

dpkedt=−iδkepke−i2Ωke(Ne,ke+Nh,ke−1)−γ⊥pke,\frac{dp_{\mathbf{k}_e}}{dt}=-i\delta_{\mathbf{k}_e}p_{\mathbf{k}_e}-\frac{i}{2}\Omega_{\mathbf{k}_e}(N_{e,\mathbf{k}_e}+N_{h,\mathbf{k}_e}-1)-\gamma_\perp p_{\mathbf{k}_e},
dNl,kedt=i2(Ωkepke∗−Ωke∗pke)+Λl,ke−γnrNl,ke−γeq(Nl,ke−fF,l,ke).(13)\frac{dN_{l,\mathbf{k}_e}}{dt}=\frac{i}{2}\left(\Omega_{\mathbf{k}_e}p_{\mathbf{k}_e}^{*}-\Omega_{\mathbf{k}_e}^{*}p_{\mathbf{k}_e}\right)+\Lambda_{l,\mathbf{k}_e}-\gamma_{nr}N_{l,\mathbf{k}_e}-\gamma_{eq}\left(N_{l,\mathbf{k}_e}-f_{F,l,\mathbf{k}_e}\right). \tag*{(13)}

where Ωke=dkeE/ℏ\Omega_{\mathbf{k}_e}=d_{\mathbf{k}_e}E/\hbar is the Rabi frequency, ll denotes electrons (ee) or holes (hh), Λl,ke\Lambda_{l,\mathbf{k}_e} describes carrier injection, γnr\gamma_{nr} is the non-radiative recombination rate, fF,l,kef_{F,l,\mathbf{k}_e} denotes the carrier quasi-equilibrium Fermi distribution, and γ⊥\gamma_\perp is the polarization decay rate. Since γ⊥\gamma_\perp is typically much larger than both the carrier and field evolution rates [30, 32], the microscopic polarization can be adiabatically eliminated by setting dpkedt=0\frac{dp_{\mathbf{k}_e}}{dt}=0. This approximation yields,

pke=−i2Ne,ke+Nh,ke−1γ⊥+iδkeΩke.(14)p_{\mathbf{k}_e}=-\frac{i}{2}\frac{N_{e,\mathbf{k}_e}+N_{h,\mathbf{k}_e}-1}{\gamma_\perp+i\delta_{\mathbf{k}_e}}\Omega_{\mathbf{k}_e}. \tag*{(14)}

The carrier density NN and the macroscopic polarization PatP_{at} are respectively defined as [30],

N=1V∑keNe,ke=1V∑keNh,ke(15)N=\frac{1}{V}\sum_{\mathbf{k}_e}N_{e,\mathbf{k}_e}=\frac{1}{V}\sum_{\mathbf{k}_e}N_{h,\mathbf{k}_e} \tag*{(15)}

and

Pat=2V∑kedkepke,(16)P_{at}=\frac{2}{V}\sum_{\mathbf{k}_e}d_{\mathbf{k}_e}p_{\mathbf{k}_e}, \tag*{(16)}

where VV is the lasing medium volume. Under the quasi-equilibrium approximation, the carrier distributions remain close to their Fermi–Dirac distributions, i.e., Nl,ke≃fF,l,keN_{l,\mathbf{k}_e}\simeq f_{F,l,\mathbf{k}_e}, allowing the resulting equations to be expressed solely in terms of the carrier density NN [30]. Substituting the adiabatic polarization (14) into the macroscopic polarization (16) leads naturally to the definition of a complex gain [30, 47],

g~(N)=ω0ε0Vℏ∑kedke2γ⊥−iδkeγ⊥2+δke2(fF,e,ke+fF,h,ke−1).(17)\tilde{g}(N)=\frac{\omega_0}{\varepsilon_0V\hbar}\sum_{\mathbf{k}_e}d_{\mathbf{k}_e}^2\frac{\gamma_\perp-i\delta_{\mathbf{k}_e}}{\gamma_\perp^2+\delta_{\mathbf{k}_e}^2}\left(f_{F,e,\mathbf{k}_e}+f_{F,h,\mathbf{k}_e}-1\right). \tag*{(17)}

whose real part corresponds to the material gain and imaginary part accounts for carrier-induced refractive-index variations. The Maxwell–Bloch equations therefore reduce to

dEdt=ivg(kp,0−kp,c)E−γc2E+12Zvgg~(N)E,\frac{dE}{dt}=i v_g(k_{p,0}-k_{p,c})E-\frac{\gamma_c}{2}E+\frac{1}{2}Zv_g\widetilde{g}(N)E,
dNdt=−vgZg(N)ngε02Zℏω0∣E∣2+JeV−γnrN,(18)\frac{dN}{dt}=-v_g Zg(N)\frac{n_g\varepsilon_0}{2Z\hbar\omega_0}|E|^2+\frac{J}{eV}-\gamma_{nr}N, \tag*{(18)}

where g(N)=Re⁡[g~(N)]g(N)=\operatorname{Re}[\widetilde{g}(N)]. Finally, defining the photon density as S=ngε02Zℏω0∣E∣2S=\frac{n_g\varepsilon_0}{2Z\hbar\omega_0}|E|^2 [46], one recovers the laser rate equations given in Eq. (2) of the main text.

Steady-state analysis

The steady-state analysis of Eqs (2) allows us to analytically evaluate the expression for the photon density threshold in the cavity:

Sth=1vggth(Λp−γnrNth),(19)S_{th}=\frac{1}{v_g g_{th}}\left(\Lambda_p-\gamma_{nr}N_{th}\right), \tag*{(19)}

showing that SthS_{th} is primarily affected by group velocity vgv_g and, to a lesser extent, by impedance ZZ through NthN_{th}, as expressed by Eq. (3). Furthermore, as expected, a lower threshold leads to a shorter time, tdt_d, to reach lasing threshold.

td=1γnrln⁡(ΛpΛp−γnrNth).(20)t_d=\frac{1}{\gamma_{nr}}\ln\left(\frac{\Lambda_p}{\Lambda_p-\gamma_{nr}N_{th}}\right). \tag*{(20)}

Finally, the laser oscillation period ωr\omega_r is affected by the impedance and the group velocity,

ωr2=Zvgg0(Λp−γnrNth),(21)\omega_r^2=\frac{Z}{v_g}g_0\left(\Lambda_p-\gamma_{nr}N_{th}\right), \tag*{(21)}

while the damping constant γ0\gamma_0 is affected only by the impedance:

γ0=Zg0gth(Λp−γnrNth).(22)\gamma_0=Z\frac{g_0}{g_{th}}\left(\Lambda_p-\gamma_{nr}N_{th}\right). \tag*{(22)}

References

  1. [1]E. M. Purcell, in Confined Electrons and Photons: New Physics and Applications, edited by E. Burstein and C. Weisbuch (Springer US, Boston, MA, 1995), pp. 839–839.
  2. [2]W. L. Barnes, S. A. R. Horsley, and W. L. V. Vos, J. Opt. 22, 073501 (2020).
  3. [3]E. P. Petrov et al., Phys. Rev. Lett. 81, 77 (1998).
  4. [4]M. Lobet et al., ACS Photonics 7, 1965 (2020).
  5. [5]A. Alù and N. Engheta, Phys. Rev. Lett. 103, 043902 (2009).
  6. [6]Y. Li and C. Argyropoulos, Opt. Express 24, 26696 (2016).
  7. [7]S. Stengel et al., APL Quantum 3, 016106 (2026).
  8. [8]I. Liberal and N. Engheta, Science Advances 2, e1600987 (2016).
  9. [9]J.-K. So et al., Appl. Phys. Lett. 117, 181104 (2020).
  10. [10]I. Liberal, A. M. Mahmoud, and N. Engheta, Nat Commun 7, 10989 (2016).
  11. [11]I. Liberal and N. Engheta, Proceedings of the National Academy of Sciences 114, 822 (2017).
  12. [12]D. I. Vulis et al., Rep. Prog. Phys. 82, 012001 (2018).
  13. [13]I. Liberal and N. Engheta, Nature Photon 11, 149 (2017).
  14. [14]N. Kinsey et al., Nat Rev Mater 4, 742 (2019).
  15. [15]L. Vertchenko, M. Nikitin, and A. Lavrinenko, J. Opt. Soc. Am. B, JOSAB 40, 1467 (2023).
  16. [16]M. Lobet et al., ACS Photonics 10, 3805 (2023).
  17. [17]M. Lobet et al., Light Sci Appl 11, 110 (2022).
  18. [18]M. H. Javani and M. I. Stockman, Phys. Rev. Lett. 117, 107404 (2016).
  19. [19]R. W. Ziolkowski, Phys. Rev. E 70, 046608 (2004).
  20. [20]H. Tang et al., Nano Lett. 21, 914 (2021).
  21. [21]Y. Li et al., Nature Photon 9, 738 (2015).
  22. [22]P. Moitra et al., Nature Photon 7, 791 (2013).
  23. [23]A. M. Mahmoud and N. Engheta, Nat Commun 5, 5638 (2014).
  24. [24]J. C. Reines et al., Opt. Express, OE 26, 21594 (2018).
  25. [25]O. Mello et al., Applied Physics Letters 120, 061105 (2022).
  26. [26]O. Mello et al., Light Sci Appl 14, 300 (2025).
  27. [27]R. Fleury and A. Alù, Phys. Rev. B 87, 201101 (2013).
  28. [28]J. Ren, S. Zhu, and Z. D. Wang, Phys. Rev. Applied 21, 044053 (2024).
  29. [29]R. Sokhoyan and H. A. Atwater, Opt. Express 21, 32279 (2013).
  30. [30]L. Lugiato, F. Prati, and M. Brambilla, Nonlinear Optical Systems (Cambridge Univ. Press, Cambridge, 2008).
  31. [31]D. Manzano, AIP Advances 10, 025106 (2020).
  32. [32]D. Meschede, Optics, Light and Lasers: The Practical Approach to Modern Aspects of Photonics and Laser Physics, 2., rev.enlarged ed (Wiley-VCH, Weinheim, 2007).
  33. [33]P. W. Milonni and G. J. Maclay, Optics Communications 228, 161 (2003).
  34. [34]T. Numai, Fundamentals of Semiconductor Lasers (Springer, New York, 2004).
  35. [35]A. Taflove and S. C. Hagness, Computational Electrodynamics: The Finite-Difference Time- Domain Method, 3rd ed (Artech house, Boston, 2005).
  36. [36]R. W. Ziolkowski and A. D. Kipple, Phys. Rev. E 68, 026615 (2003).
  37. [37]R. W. Ziolkowski, Phys. Rev. E 63, 046604 (2001).
  38. [38]S.-H. Chang and A. Taflove, Opt. Express, OE 12, 3827 (2004).
  39. [39]X. Jiang and C. M. Soukoulis, Phys. Rev. Lett. 85, 70 (2000).
  40. [40]R. Contractor et al., Nature 608, 692 (2022).
  41. [41]X. Huang et al., Nature Mater 10, 582 (2011).
  42. [42]A. Debaq and M. Lobet, (2026).
  43. [43]M. Lyubavov et al., Science 377, 425 (2022).
  44. [44]J. B. Pendry, Opt. Express, OE 31, 452 (2023).
  45. [45]S. A. R. Horsley and J. B. Pendry, Proc. Natl. Acad. Sci. U.S.A. 120, e2302652120 (2023).
  46. [46]J. D. Jackson, Classical Electrodynamics, 3. ed., [Nachdr.] (Wiley, Hoboken, NY, 2009).
  47. [47]T. Takahashi, M. Nishioka, and Y. Arakawa, Appl. Phys. Lett. 58, 4 (1991).

Paper details

vlnc:W_vsbqmk9z
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 2

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

  • Physics and Astronomy

Generative AI tools

View statement
Looking for related papers…
Contents