Combs, fast and slow: non-adiabatic mean field theory of active cavities
David Burghoff · 2024
Read Combs, fast and slow: non-adiabatic mean field theory of active cavities by David Burghoff (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2411-03281.txt.
Description
Abstract Integrated frequency combs based on active cavities are of interest for a wide range of applications. An elegant description of these cavities is based on mean-field theory, which averages the effect of internal dynamics occurring within a round trip.
Catalog record
| CREATOR | David Burghoff |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 98 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2411-03281
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Combs, fast and slow: non-adiabatic mean field theory of active cavities David Burghoff Abstract Integrated frequency combs based on active cavities are of interest for a wide range of applications. An elegant description of these cavities is based on mean-field theory, which averages the effect of internal dynamics occurring within a round trip. Lasers based on media with slow gain dynamics can be described by solving the population over many round trips, while lasers based on fast gain media can be described by adiabatic elimination. However, most gain media actually have both fast and slow components, and effects often ascribed to fast gain media are known to arise even in slower gain media. Here, we develop an operator-based mean-field theory that non-adiabatically describes the dynamics of bidirectional active cavities, both fast and slow. It is based on first principles and semi-exactly replaces the Maxwell-Bloch equations, but is flexible enough to accomodate non-trivial lineshapes and population dynamics. As an example, we use this formalism to establish an additional constraint on the formation of frequency-modulated combs. Our results are general and apply to any bidirectional or unidirectional active cavity, and as a result, generalize to essentially any chip-scale laser. pacs Valid PACS appear here I Introduction Lasers are some of the richest dynamical systems to have ever been produced and still produce surprises. Over the last few years, a number of comb states have been shown to appear in systems with fast gain media whose population dynamics occur much faster than their round-trip time. Perhaps chief among these is the linearly-chirped frequency-modulated (FM) comb state, which was first conclusively identified in quantum cascade lasers using SWIFTS [ 1 , 2 ] , but has also been shown to appear in a wide range of semiconductor systems [ 3 , 4 , 5 , 6 , 7 ] using a variety of experimental techniques [ 8 , 9 , 10 ] . Other comb states relying on the fast nature of the gain medium include the quantum walk state [ 11 ] , the Nozaki-Bekki hole [ 12 ] , and the production of bright-dark pairs in coupled rings [ 13 ] . When the gain medium’s population dynamics are fast relative to the round trip time, the population and coherence dynamics can be adiabatically eliminated [ 14 ] , permitting analytical solutions to be found [ 15 ] . This is in stark contrast to the usual description of slow gain media, which are typically described by the Haus master equations that eliminate the dynamics entirely [ 16 ] . But how fast is fast? The most complete description of lasers above threshold are the Maxwell-Bloch equations. These are completely exact and can be solved numerically, but usually cannot be solved analytically except in the case of very specific trivial solutions. Moreover, the parameter space is large, and the specific gain dynamics of the laser must be included—meaning one is subjected to an enormous number of parameters that in many cases are not even measurable. It is for this reason that novel states are mostly driven by experiment. For example, even though the linearly-chirped FM state is now recognized as a somewhat universal state in semiconductor lasers, it had not been predicted at all. A more recent development has been the creation of mean-field theories that are analogous to the famed Lugiato-Lefever equation [ 17 ] . When the internal dynamics of a field within a round-trip are small, these changes can be averaged and ignored. The Lugiato-Lefever equation has been known for several years to describe Kerr combs effectively [ 18 , 19 ] , but was not applied semiconductor lasers since they typically have large outcoupling losses (violating the main assumption of mean-field theory). However, we showed in [ 15 ] that by normalizing the field to its continuous-wave (CW) profile, one could construct a mean-field theory that applies to lasers—essentially the LLE for active cavities. When applied to a bidirectional Fabry-Perot cavity it can result in dynamics governed by a phase-driven nonlinear Schrodinger equation, and when applied to a unidirectional cavity it becomes a complex Ginsburg-Landau equation [ 20 ] . The chief technique that has been used to describe fast gain media analytically has been the adiabatic approximation [ 14 , 15 , 21 , 12 , 11 ] . One can show from the Bloch equations that the population and coherence are each related to the quantity (1+T_{i}\partial_{t})^{-1} , where T_{1} is the population inversion lifetime and T_{2} the coherence lifetime. In the adiabatic approximation, the population inversion and coherence are eliminated by taking the low-frequency limit, computing (1+T_{i}\partial_{t})^{-1}\ \approx 1-T_{i}\partial_{t}+... , and only taking the first term or two. While adequate in the limit of very fast gain media, it has some troubling aspects. The first is that the geometric series intrinsic to the adiabatic approximation is not convergent for \omega T_{i}>1 . Thus, frequencies too far from the carrier frequency cannot possibly be well-described by it. Secondly, it is numerically poorly-behaved. Such a series generates high frequency components that can cause numerical divergence. As a result, it cannot be applied to systems whose bandwidth is much greater than 1/T_{1} or 1/T_{2} . Here, we exactly solve the Maxwell-Bloch equations in the non-adiabatic limit, producing both exact normalized master equations and a semi-exact mean-field theory. By introducing Lorentzian operators, we rigorously show that the dynamics of an arbitrary bidirectional cavity can be solved using either \displaystyle\frac{1}{v_{g}}\frac{\partial f}{\partial t}+\frac{\partial f}{\partial z} \displaystyle=\frac{1}{2}\hat{g}_{\textrm{NL}}f\quad\text{or} (1) \displaystyle\frac{1}{v_{g}}\frac{\partial f}{\partial T} \displaystyle=\frac{1}{2}\langle\hat{g}_{\textrm{NL}}\rangle f, (2) where the first equation represents an exact master equation in the normalized field f in terms of the nonlinear gain \hat{g}_{\textrm{NL}} and the group velocity v_{g} , and the latter represents a semi-exact mean-field theory. Either approach that can be solved extremely efficiently using standard split-step methods. Moreover, because our approach is operator-based, different lineshapes and different population dynamics can be included without any additional analysis. With this approach, we demonstrate an additional constraint on FM combs not contained within the adiabatic extendon theory—that there is a maximum limit to the gain recovery time in which extendons cannot form. We also introduce multi-component mean-field theory, which allows for the parsimonious modeling of gain media with both fast and slow components. II Exact master equations and mean-field theory The detailed first-principles derivation is given in the appendix, and here we focus on the key results. First, we use the Maxwell-Bloch equations to derive our exact theory. Starting from the evolution of a two-level system with a population lifetime T_{1} and a coherence lifetime T_{2} , we introduce the operators \displaystyle\hat{L}_{i}\equiv\left(1+T_{i}\frac{\partial}{\partial t}\right)^{-1} (3) in order to capture all of the long-term and short-term dynamics. These can be intepreted as Lorentzians in frequency space, since L_{i}(\omega)=(1+i\omega T_{i})^{-1} . Using this, we then show in Appendices A and B that a field E_{f}=Ae^{i\omega_{0}t}+A^{*}e^{-i\omega_{0}t} gives rise to a nonlinear gain of the form \displaystyle\hat{g}_{\text{NL}}=g_{0}\hat{L}_{2}\left[1+\frac{1}{P_{s}}\operatorname{Re}\hat{L}_{1}A^{*}\hat{L}_{2}A\right]^{-1} (4) where P_{s}\equiv E_{s}^{2} is the saturation field squared and g_{0} is the small-signal gain. When the usual limits are taken for slow gain, \hat{L}_{2}\rightarrow L_{2}(\omega) and \hat{L}_{1}\rightarrow\hat{I} , so one recovers the standard form of gain saturation in terms of a homogeneously-broadened lineshape. Still, this expression clarifies that