![Controlling single-mode random laser specifications
a, Illustrations of an approach to control the wavelength and size of the single random laser mode. The disorder configuration and device size remain the same. b, Colour plot of the single-mode random laser wavelength versus air hole size (r0/a) and disorder strength (γ). c,d, μPL spectra from selected devices along line (1) and (2) of b: r0/a varying (c) and γ varying (d). e, Relationship between random laser wavelength and dimensionless structural parameter z=-r¯0∕a+βγ¯\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{z}} = - \left[ {\left( {\bar r_0/a} \right) + \beta \bar \gamma } \right]$$\end{document} for selected values of fitting parameter β. f, μPL spectra and NSOM images (insets) for selected devices along line (3) of b: λL fixed. g,h, Lasing wavelength (g) and effective modal width (h) along line (3) of b, versus γ. i, μPL spectra and NSOM images (insets) for selected devices along line (4) of b: weff fixed. j,k, Lasing wavelength (j) and effective modal width (k) along line (4) of b, versus r0/a.](https://www.researchgate.net/publication/332279695/figure/fig3/AS:745641600892933@1554786219642/Controlling-single-mode-random-laser-specifications-a-Illustrations-of-an-approach-to_Q320.jpg)
A preview of this full-text is provided by Springer Nature.
Content available from Nature Photonics
This content is subject to copyright. Terms and conditions apply.
Letters
https://doi.org/10.1038/s41566-019-0407-5
1Department of Physics and Astronomy, Seoul National University, Seoul, Republic of Korea. 2Inter-university Semiconductor Research Center, Seoul
National University, Seoul, Republic of Korea. 3Université de Lyon, Institut des Nanotechnologies de Lyon-INL, UMR CNRS 5270, CNRS, Ecole Centrale de
Lyon, Ecully, France. 4Institute of Applied Physics, Seoul National University, Seoul, Republic of Korea. *e-mail: hsjeon@snu.ac.kr
Random lasers1 are fascinating devices due to the absence of a
conventional cavity structure and their counterintuitive lasing
mechanism. However, they are also notorious for their unpre-
dictability. Despite their many unusual properties2–5, random
lasers are unlikely to achieve the ubiquitous acceptance of
conventional lasers unless the underlying lasing mechanisms
that govern their operation are thoroughly understood and
their exotic properties are appropriately regulated. Recent
demonstrations of localized random lasers are considered a
breakthrough in the field because structural disorders were
engineered in a top-down manner6–8. Nevertheless, the ori-
gin of the lasing phenomenon and the controllability of these
devices have not been adequately addressed. Lately, we have
experimentally proven that photonic band-tail eigenstates—
manifestations of photonic Anderson localizations—are
responsible for random lasing in a compositionally disordered
photonic crystal platform9. Herein, we demonstrate that the
process of governing the band-tail states offers a unique
opportunity to finally regulate random lasers.
Ever since their first demonstration10, random lasers have
attracted tremendous attention because they obviate the need for a
cavity structure. Ironically, however, this convenient and advanta-
geous property is an intrinsic obstacle that has hindered the fur-
ther development of random lasers. Their randomness or disorder
typically results in very obscure and cumbersome processes when
systematic approaches are undertaken in their structural design and
the analysis of their results. Two types of random laser have been
demonstrated so far: diffusive random lasers10–12 (the first version
of random lasers, which are prepared from diffusive random media
in a bottom-up approach and thus are inherently uncontrollable)
and localized random lasers (based on disordered photonic crystal
structures)6–8. It has been claimed6 that the origin of lasing in the
latter is photonic Anderson localization13,14. Moreover, Anderson-
localized modes were suggested as an ideal platform for random
lasing action15–18. Nevertheless, no attempt has been made to tailor
the properties of random lasers due in part to the absence of an
appropriate method to control the random structures and/or a deep
understanding of the lasing mechanism. However, we note a few
attempts based on post-tuning methods, such as tuning the excita-
tion beam profile19, external fields20 or temperature21, but not on a
pre-designing approach at a fundamental level.
Quite recently, we proved experimentally that photonic crystal
alloys can induce photonic band-tail (PBT) states (inside a bandgap
but near the band-edges)9, which are counterparts to the electronic
band-tail states in heavily doped semiconductors22. We also showed
that some PBT states are the photonic Anderson localization modes,
which not only satisfy the Ioffe–Regel criterion (in a much relaxed
form)23 but also possess cavity Q factors that are adequate to facilitate
lasing when sufficient optical gain is provided. Naturally, our next
research objective is to engineer photonic crystal alloy structures
to manipulate the PBT states and thus the properties of the ran-
dom laser. The key to accomplishing this challenging goal lies in the
compositional disorder, a structural scheme by which our photonic
crystal alloy system is constituted (Fig. 1a). Note that a photonic
crystal lattice structure is retained while disorder is incorporated
via controlled fluctuations of the effective refractive index. Distinct
from the lattice disorder utilized exclusively by other random laser
researchers1–8,10–12,15–21, the compositional disorder is a versatile and
powerful approach for introducing and controlling disorder9,24,25.
Our random lasers, which we have named band-tail lasers based
on the origin of the lasing action, were fabricated using a thin-wave-
guide slab containing InAsP/InP multiple quantum wells (MQWs)
(see Methods). For each device, an array of air holes was perforated
into the waveguide slab to construct a two-dimensional (2D) com-
positionally disordered photonic crystal structure. The controlled
fluctuations in the effective refractive index are implemented via
the differentiation of photonic atoms, that is, by varying the sizes of
the air holes (Fig. 1b). A compositionally disordered photonic crys-
tal platform is therefore associated with a photonic band structure,
regardless of the degree and configuration of the disorder, because
of the retention of its crystallinity9,24,25. This is in contrast to previous
localized random lasers based on lattice disorder6–8, for which nei-
ther a photonic crystal structure nor the corresponding band struc-
ture are conceivable unless the disorder strength is weak.
We chose the hexagonal lattice as our photonic crystal backbone
structure because the associated bandgap is wide enough to cover the
entire emission band of the MQWs. Each device is patterned into a
hexagon, commensurate with hexagonal crystalline symmetry. The
device size can then be quantified by the normalized side length of
the hexagon L/a (Fig. 1a), where a = 450 nm is the photonic crystal
lattice constant, which is tuned to the MQW emission. The multi-
plicity of photonic atoms was chosen to be M = 4, so our composi-
tionally disordered photonic crystal platform is thus a quaternary
alloy system. The configurational degree of freedom (DOF) is then
given by an exponential function: [DOF] = Mn = 4n, where n is the
number of photonic crystal lattice sites in a device. For a device size
of L/a = 3, for which n = 37, the corresponding DOF is quite large:
[DOF] = 437 ≈ 1022. We followed the scheme proposed by Conti and
colleagues26 in assigning the air-hole sizes: r = r0(1 + γξ), where r0 is
the reference radius, γ (≥0) is the degree of disorder, and ξ is the
deviation, which varies uniformly within the range of –1/2 ≤ ξ ≤ 1/2.
For our quaternary system, ξ = (−1/2, −1/6, + 1/6, +1/2).
Our strategy for controlling the PBT-based random lasers con-
sists of two steps. The first involves the identification of optical
states existing within a given disordered photonic crystal pattern
and the subsequent selection of a specific Anderson-localized PBT
mode for a single-mode lasing action. The details of the pattern
are then engineered (through r0 and γ) to achieve tailored laser
Taming of random lasers
Myungjae Lee 1,2, Ségolène Callard 3, Christian Seassal3 and Heonsu Jeon 1,2,4*
NATURE PHOTONICS | VOL 13 | JULY 2019 | 445–448 | www.nature.com/naturephotonics 445
The Nature trademark is a registered trademark of Springer Nature Limited.
