Manipulating strong coupling between exciton and quasi-bound states in the

continuum resonance via generalized temporal coupled mode theory

Meibao Qin,1 Junyi Duan,2, 3 Shuyuan Xiao,2, 3, ∗ Wenxing Liu,1 Tianbao Yu,1, † Tongbiao Wang,1 and Qinghua Liao1
1
Department of Physics, Nanchang University, Nanchang 330031, People’s Republic of China
2
Institute for Advanced Study, Nanchang University, Nanchang 330031, People’s Republic of China
3
Jiangxi Key Laboratory for Microscale Interdisciplinary Study,
Nanchang University, Nanchang 330031, People’s Republic of China
Strong coupling exhibits unique ability to preserve quantum sates between light and matter,
which is essential for the development of quantum information technology. To explore the physical
arXiv:2111.00689v2 [physics.optics] 15 Dec 2021

mechanism behind this phenomenon, we employ the tight-binding model for expanding the temporal
coupled-mode theory, with the absorption spectrum formula of coupled system directly obtained in
an analytical way. It reveals all the physical meaning of parameters defined in our theory, and shows
how to tailor lineshapes of the coupled systems. Here, we set an example to manipulate the strong
coupling in a hybrid structure composed of excitons in monolayer WS2 and quasi-bound states in
the continuum supported by the TiO2 nanodisks. The simulated results show that a clear spectral
splitting appeared in the absorption spectrum, which can be controlled by adjusting the asymmetric
parameter of the nanodisk metasurfaces and well fitted through our theoretical predictions. Our
work not only gives a more comprehensive understanding of such coupled systems, but also offers
a promising strategy in controlling the strong light-matter coupling to meet diversified application
requests.

I. INTRODUCTION and robust existence14–17 .

In recent years, many researchers have investi-
Enhancing the light-matter interaction could offer a gated strong coupling between the TMDs and plas-
large variety of applications in many areas, such as quan- monic/dielectric nanostructures18–35 , and developed a
tum electrodynamics, nonlinear optics, photonics, and general approach to employ the coupled harmonic oscil-
so on1,2 . Strong coupling occurs when the coherent ex- lator model for such hybrid systems, in which nanostruc-
change rate between the two subsystems with the coin- tures are treated as an isolated system and the influ-
cide frequencies exceeds their individual decay rate, and ence of external excitation is ignored. However, the cou-
results in the formation of a new hybrid light-matter pled systems are open in nature, where different resonant
state3–5 . A key aspect of achieving this lies in enhanc- modes couple not only with each other via near field but
ing the coupling strength g, which can be determined as also with the external free space36 . Therefore, an analyt-
g = µ · E ∝ V1 in the dipole approximation6. Here, µ ical theory that can consider the near field and external
and E are the transition dipole moment and the local free space in the same foot is in urgent need. As an early
electric field intensity, respectively. In order to fulfill attempt, the existing temporal coupled mode theory can
this requirement, monolayer transition-metal dichalco- deal with the coupling between propagation modes and
genides (TMDs) emerge as a kind of promising materials background modes, unfortunately, it cannot resolve the
due to their large transition dipole moment and other problem that is bound states coupled to various ports,
extraordinary electronic and optical properties, such as such as exciton modes in strongly coupled systems. On
high binding energy owing to the quantum confinement this issue, the tight-binding method that can divide the
of the atomic and direct band gaps7,8 . One of the feasible wave function into propagating state and bound state,
ways to enhance the strong coupling is to integrate the becomes an ideal platform to solve such bound states.
monolayer TMDs to the confined electromagnetic field. Therefore, utilizing the tight-binding method for gener-
In this case, optical nanocavities exhibit excellent light alizing temporal coupled mode theory is expected to be
trapping, and have been proven to be an efficient way to helpful to deal with the interaction among each resonant
realize strong field enhancement. The conventional plas- modes and with the free space as well.
monic nanostructures confine light within subwavelength In the present work, we derive a formal theoretical
mode volume but suffer heavy ohmic loss in the visible work based on the photonic tight-binding method and
region, which limits to their practical applications9,10 . In temporal coupled mode theory, to better handle the inter-
contrast, dielectric nanostructures do have low intrinsic action between coupled system and external free space.
loss, and allow the existence of the multiple interference As a proof of demonstration, we investigate a hybrid
in resonant high-index dielectric nanostructures, offering structure composed of monolayer WS2 placed on top of
novel opportunities for the controlling of the light-matter TiO2 nanodisk metasurfaces supporting magnetic dipole
interaction11–13 . In particular, a recently emerged con- quasi-BIC resonance. In order to fulfill strong coupling
cept of bound states in the continuum (BIC) enables a condition, we keep the frequencies of the two coupled
promising way to achieve strong light-matter interaction, modes consistent by adjusting the height and the asym-
which benefits from their better confinement of energy metric parameter of the nanodisk metasurfaces, and the
 2

couple strength g = 17.91 meV can be achieved through In order to apply the tight-binding method to manip-
benchmark calculations. The typical anti-crossing be- ulate strong coupling, we need to construct a coupling
havior of the Rabi splitting 24.04 meV in the absorp- matrix D̂ between the resonant modes and the ports. In
tion spectrum can be observed, which agrees well with the proposed hybrid system, only the resonator exhibits
our proposed theory. Moreover, we applied the approach loss rate γ, and one incident port means q = 1 in the Eq.
to tailor the absorption spectrum in the various regimes (4), so
√the coupling
 matrix can be expressed by the form
from weak coupling to strong coupling. Our work pro- γ 0
vides a broad and deep insight into the properties of the D̂= √ . The coupling among eigen modes and the
γ 0
coupled systems, which is much closer to the actual sit- loss of the system can be expressed as
uation, and holds a great potential for controlling strong    
coupling interaction. ωn 0 0 g
Ω̂ = Ω̂1 + Ω̂2 = + , (7)
0 ωe g 0
D̂† D̂
II. GENERALIZED TEMPORAL COUPLED Γ̂ = Γ̂rad + Γ̂abs = + Γ̂abs
MODE THEORY    2
γ 0 0 0
= + , (8)
0 0 0 δ
We develop the theory on the basis of the tight-binding
method37,38 , to provide insight into the propagating state where Ω̂ is a 2 × 2 matrix determined by the interaction
and bound state of a coupled system consisted of exci- of the eigenmodes. ωn and ωe represent the resonance
ton in monolayer TMDs and arbitrary resonator under frequency of the resonator and exciton, respectively. δ
certain external illumination. The Hamiltonian of such is the absorption loss of exciton and g gives the cou-
open system can be written as pling strength between resonance mode and exciton. Γ
describes the loss of the hybrid system, related by the
Ĥ = Ĥb + Vˆn + Vˆe , (1) radiation and absorption loss rates.
Based on the above derivations, we can generalize the
where Ĥb is the Hamiltonian of the background, V̂n and temporal coupled mode theory39 , to describe the light-
V̂e represent the potential contributed by the resonator matter interaction in the coupled system,
and exciton, respectively. Then, we need to solve the  
d an

ωn + iγ g
 
an
following eigen equation of the whole system =i (9)
dt ae g ωe + iδ ae
√ √   + 
Ĥ|Ψi = ω|Ψi, (2) γ γ s1
+ ,
0 0 0
here |Ψi = |Ψin i + |Ψrs i is the total wave function with
Ψin and Ψrs describing the incident port and resonance |s− i = Ĉ|s+ i + D̂|ai, (10)
mode, respectively. For this open system with two reso- Here, |s+ i and |s− i are the amplitude of plane wave in
nance modes, Ψin and Ψrs can be further written as the input and output, respectively. Ĉ is the direct scat-
|Ψrs i = an |Ψn i + ae |Ψe i, (3) tering matrix with reflection and transmission between
X incoming and outgoing waves. |ai is the resonance am-
|Ψin i = s+
q |Ψq i, (4) plitude of the mode with a general time dependence of
q ∝ eiωt , the dynamic equations for resonance amplitudes
can be written in the following form:
where an and ae are the strength of the electric field ex-
cited by the resonator and exciton, respectively, under ˆ − Ω̂) − iΓ̂]−1 D̂T |s+ i,
i|ai = [(Iω (11)
external illumination described by s+q denoting the ex- ˆ − Ω̂) + iΓ]−1 .
ha|i = −hs+ |(D̂T )† [(Iω (12)
citation amplitudes at different incoming ports. Due to
unavoidable radiation in open systems, we should not According to the energy conservation, we have the rela-
only consider near field (NF), but also include the far tion for resonance amplitude and time,
field (FF) propagating to the external continuum. |Ψn i
d|a|2 d ha | ai
and |Ψe i can be expressed as =
dt  dt 
|Ψn i = |ΨNF FF
n i + |Ψn i, (5) = ha| ∂t† |ai + ha| ∂t |ai
NF
|Ψe i = |Ψe i + |ΨFF
e i, (6)
= −2 ha| Γ̂ |ai , (13)
with |ΨNF
n i |ΨFF
and n i representing the NF and FF of the and the absorption of the coupled systems at the fre-
resonator wave function, respectively. |ΨNF FF
e i and |Ψe i quency can be finally obtained,
describe the NF and FF of the exciton wave function,
ˆ
hs+ |(D̂T )† [(Iω− ˆ
Ω̂)+iΓ̂]−1 Γ̂abs [(Iω− Ω̂)−iΓ̂]−1 D̂T |s+ i
respectively. It is noted that the exciton mode in high A= hs+ |s+ i
refractive dielectric is a kind of bound sate. Hence the
2δγg2
far field of its wave function can be defined as |ΨF F
e i = 0.
= [(ω−ωn )(ω−ωe )−δγ−g2 ]2 +[δ(ω−ωn )+γ(ω−ωe )]2 . (14)
 3

In such a formula, the absorption spectrum can be di- pling.

rectly predicted, and will exhibit an asymmetric Fano The full-wave numerical simulations are performed
line shape with two peaks and one dip, when satisfying with the finite element method (FEM) via commercially
2 2
the condition g 2 > γ +δ2 > δγ. Following this, strong available software COMSOL Multiphysics. In the follow-
coupling can be realized by the control of the radiation ing simulation calculations, the y-polarized plane wave
loss and absorption loss. is normally incident along the z direction, where the pe-
riodic boundary conditions are adopted in both the x
and y directions, and the perfectly matched layers are
III. STRUCTURE DESIGN AND NUMERICAL employed in the z direction.
MODEL

In order to identify the reliability of the generalized IV. RESULTS AND DISCUSSIONS
temporal coupled mode theory, we consider an absorp-
tion system composed of a monolayer WS2 laid flat on
To engineer the radiation loss in the proposed nanodisk
an array of TiO2 nanodisk with a SiO2 substrate, as the
metasurfaces, we first analyze the resonant mode gov-
sketches shown in Fig. 1(a). Each unit cell is charac-
erned by the BIC physics without the presence of WS2 .
terized by a square lattice with a period P = 410 nm,
Fig. 2 (a) presents the simulated transmission spectrum
height H = 85 nm, radius R = 140 nm, and a nanohole
(blue points) with a radius of the off-centered hole r = 50
with a variable radius is introduced at a fixed distance
nm, i.e., the asymmetric parameter α = 0.127, show-
d = 70 nm away from the center of the nanodisk. The
πr 2
ing an asymmetric line-shape Fano resonance with a dip
perturbation with an asymmetric parameter α = πR 2 is around 616 nm, and the simulation result is well fitted
thus introduced into the in-plane symmetric of the struc- (red line) by the Fano formula39,46,47 , which suggests that
ture, which can open a radiation channel and transform the quasi-BIC obtains an energy exchange with the con-
a symmetric-protected BIC into a quasi-BIC16,40,41 . tinuum free-space radiation modes. The near field dis-
tribution is plotted in the inset of Fig. 2 (a), where
the bare metasurfaces hold stronger excitation of elec-
tric field at the center of the off-centered hole, and the
distinct circular displacement current in the x-y plane of
the nanodisk suggests that the energy is strongly local-
ized by the magnetic dipole moment oscillation along the
z direction48 . In order to further identify the magnetic
dipole quasi-BIC resonance, the electromagnetic multi-
poles are decomposed under the Cartesian coordinates
in Fig. 2(b), and it is obvious that radiation contribu-
tion of the magnetic dipole ultimately dominates.

FIG. 1. (a) The schematic of the proposed coupled system

composed of a monolayer WS2 deposited on the nanodisk
metasurfaces. (b) The dielectric constant of the monolayer
WS2 . The blue and red lines represent the real and imagi-
nary parts, respectively.

For the material modelling, TiO2 can be a good can-

didate owing to its high refractive index nTiO2 = 2.6 and
negligible extinction coefficient in the range of the visi-
ble wavelengths, which can constitute optically-induced
electric and magnetic Mie resonators and strongly
confine electromagnetic energy within the structured
FIG. 2. (a) The simulated and theoretical transmission spec-
metasurfaces42–44 . The index of SiO2 is utilized as the tra of nanodisk metasurfaces with the asymmetric parameter
substrate and assumed with a real constant refractive in- α = 0.127. The inset shows the corresponding electric field
dex nSiO2 =1.45. For WS2 , a Lorentz oscillator model distribution of |Exy | at the resonance, overlaid with arrows
with the thickness of 0.618 nm is adopted from the ex- indicating the direction of the displacement current. (b) The
perimental parameters by Li et al45 , as depicted in Fig. scattered power of the multipole moments under the Carte-
1(b). The most interesting finding is that the imaginary sian cordinates, including the electric dipole (ED), magnetic
part appears a sharp peak (red line) around the transi- dipole (MD), toroidal dipole (TD), electric quadrupole (EQ),
tion energy of the exciton 2.01 eV. Therefore, WS2 is a and the magnetic quadrupole (MQ) scaled in the logarithmic
favourable option to realize the strong light-matter cou- plot.
 4

Next, we turn to the case where a monolayer WS2 with oretical values predicted by our theoretical frame work.
a dissipate loss rate δ is spread on the nanodisk metasur-
faces. As shown in Fig. 3(a), a clear spectral splitting
with two peaks (P1 and P2 ) and one dip (D) is observed
in the visible range from 600 nm to 630 nm. The simu-
lated absorption spectrum (blue points) can be well fitted
(red line) by Eq. (14) via the generalized temporal cou-
pled mode theory. The coupling strength and Rabi split-
ting are found as 17.91 meV and 24.04 meV, respectively,
from the fitting parameters in Fig. 3(a). The absorption
loss δ = 11 meV was obtained from data by Li et al45 ,
and the radiation rate γ = 7.42 meV can be extracted
√
δ 2 +γ 2
from Fig. 2(a). Significantly, g > |δ−γ|
2 and g > 2 ,
which indicate that the strong coupling between the mag-
netic dipole quasi-BIC resonance and exciton is achieved. FIG. 4. (a) The absorption spectra of the new hybrid state
In addition, the electric field distributions of the peaks with different heights H and the asymmetric parameter α =
and dip are compared in Fig. 3(b)-(d), and the electric 0.127. (b) The red line and blue line depict the individual
fied enhancement of peaks P1 (614.3 nm) and P2 (621.7 exciton and quasi-BIC mode, and the orange stars and blue
nm) are found to be much higher than that of the dip D stars are the upper polariton branch (UPB) and lower polari-
(617.4 nm). It dramatically differs from the individual ton branch (LPB) sates, respectively.
exciton or the quasi-BIC mode, since the original energy
levels are hybridized and formed two new energy levels Fig. 5(a) shows the simulated transimission spec-
once the strong light-matter coupling condition is satis- tra with different asymmetric parameters, and the res-
fied. onant wavelengths are all set around 616 nm by tuning
the height. When the asymmetric parameter starts at
α = 0.045, the spectral line width is very narrow (only 7.8
meV), and the local field intensity reaches over 30 times
more than the incident field. As α gradually increases to
0.127, the spectral line width is 15 meV, the local field
only increases by 12 times. Moreover, when α is fur-
ther increased to 0.215, the local field density decreases
sharply with the line width increasing even broader. It
might be intuitively that the giant strong coupling can
be realized by the larger electric field enhancement with
the narrower spectral line width, however, the Rabi split-
ting behaves distinctly from such consideration. Fig.
5(b) displays the evolution of the absorption line shape
with the asymmetric parameter under the energy detun-
ing ωB − ωE = 0. When the asymmetric parameter is
α < 0.005, there is only one peak near 616 nm, which is
FIG. 3. (a) The simulated and theoretical absorption spectra similar to the absorption of uncoupled exciton, suggest-
of the coupled system with the asymmetric parameter α = ing that the excited state is only slightly perturbed by
0.127. (b)-(d) The normlized electric field distributions of the the quasi-BIC. As α proceeds from 0.045 to 0.127, the
new hybrid states at absorption peaks labeled by P1 and P2 single peak evolves into two peaks. Finally, when α fur-
and at the absorption dip marked by D, respectively.
ther exceeds 0.215, the absorption spectrum becomes a
Lorentz-line shape. Further comparison between the val-
Fig. 4(a) presents the simulated optical response of ues of Rabi splitting in the strong coupling regime shows
the coupled system with different heights. A typical anti- that the asymmetric parameter is α = 0.045 with the
crossing behaviour can be captured in the 2D color di- Rabi splitting 18.25 meV and the coupling strength 11
agram, which further confirms that the strong coupling meV, while asymmetric parameter α = 0.127, the Rabi
between the magnetic dipole quasi-BIC and exciton is splitting is 24.04 meV and the coupling strength is 17.91
reached, and significant upper polariton branch and lower meV. This unsuspected behavior can be explained by the
polariton branch sates in the hybrid system are finally fact that the coupling effect is simultaneously affected by
formed. At zero detuning ωB − ωE = 0, where the ωB the spectral line width and electric filed enhancement.
and the ωE are resonance frequency of the quasi-BIC res- Despite the local electric field is stronger for a smaller α,
onator and exciton, respectively, the Rabi splitting 24.04 the spectral line width of the quasi-BIC resonance mode
meV can be extracted by the numerical simulation results is more narrower, when it is far less than the exciton line
shown in Fig. 4(b), which is matched well with the the- width of 22 meV, the number of photons that can partici-
 5

pate in the coupling effect is much smaller which degrades nipulated by simply controlling the radiation loss. For
the Rabi splitting. When α is further increased to 0.215, example, when the radiation loss is very small, the Rabi
the local field density decreases sharply, and the number splitting will gradually disappears and the coupled sys-
of photons supplied is not enough. Hence, Rabi splitting tem enters the weak coupling region, in which the Purcell
significantly increases with the number of photons that effect can be enhanced for the study of emitters6,49 . As
can participate in the coupling, and the balance between the radiation loss increases to supply enough photons,
the line width and the local electric field enhancement of and the spectrum splits into two peaks and achieves the
quasi-BIC mode needs to be carefully considered when strong coupling, which can be used for quantum electro-
designing the strong coupling system. dynamics, nonlinear enhancement, and so on50–52 . How-
ever, when the radiation loss satisfies the critical con-
dition g 2 = δγ, the split spectrum evolves into Lorentz
lines which can be used to enhance the absorption of 2D
materials41,53–56 .

V. CONCLUSIONS

In conclusion, we have generalized the temporal cou-

pled mode theory to study the strong light-matter cou-
pling, in which all involved parameters can be directly
computed by our fitting procedures. In an exemplary
system composed of monolayer WS2 on TiO2 nanodisk
FIG. 5. (a) The simulated transimission spectra with dif-
ferent asymmetric parameters α. (b) The evolution of the
metasurfaces, the strong coupling between the magnetic
absorption line shape with the asymmetric parameter under dipole quasi-BIC resonance and exciton is realized with
the energy detuning ωB − ωE = 0. the coupling strength up to 17.91 meV. Furthermore, it
is important to balance the linewidth of the quasi-BIC
To further explore the physical mechanism behind the mode and local electric field enhancement since both of
them affect the strong coupling. Beyond this example,
above phenomenon, we deduce the Eq. (14) under the
detuning condition, ωB − ωE = 0 and the absorption of our proposed analytical model has clear physical mean-
ing, high fitting accuracy, and can be widely applied
the hybrid system can be further expressed as
to various coupled systems composed of different reso-
2δγg 2 nant nanostructures and exciton in TMDs throughout
A= . (15) the spectrum, which is potential for future compact quan-
[δγ + g 2 ]2
tum information devices.
It can be seen from the formula that there are only two
factors that affect the absorption spectrum, the radia-
tion loss and coupling grength, since the absorption loss ACKNOWLEDGMENTS
of exciton is constant in the design. It is known that the
coupling strength g is proportional to the electric field in- This work is supported by the National Natural Sci-
tensity of the interface between the two-dimensional ma- ence Foundation of China (Grants No. 11947065, No.
terials and the metasurfaces29 . As mentioned earlier, the 12064025), the Natural Science Foundation of Jiangxi
magnetic dipole moment oscillation along the z direction, Province (Grant No. 20202BAB211007), the Interdisci-
and the material is tiled on the surface of the metasur- plinary Innovation Fund of Nanchang University (Grant
faces in our proposed model, so the electric field strength No. 20199166-27060003), and the Major Discipline
of the contact surface should be related to the local field Academic and Technical Leaders Training Program of
strength and radiation loss of the quasi-BIC resonance. Jiangxi Province.
It suggests that the strong coupling system can be ma- M.Q. and J.D. contributed equally to this work.

∗ 4
[email protected] P. Torma and W. L. Barnes,
†
[email protected] Rep. Progr. Phys. 78, 013901 (2014).
1 5
N. I. Zheludev and Y. S. Kivshar, D. S. Dovzhenko, S. V. Ryabchuk, Y. P. Rakovich, and
Nat. Mater. 11, 917 (2012). I. R. Nabiev, Nanoscale 10, 3589 (2018).
2 6
S. Xiao, T. Wang, T. Liu, C. Zhou, X. Jiang, and J. Zhang, J. T. Hugall, A. Singh, and N. F. van Hulst,
J. Phys. D Appl. Phys. 53, 503002 (2020). ACS Photonics 5, 43 (2018).
3 7
H. Deng, H. Haug, and Y. Yamamoto, F. Xia, H. Wang, D. Xiao, M. Dubey, and A. Ramasub-
Rev. Mod. Phys. 82, 1489 (2010). ramaniam, Nat. Photonics 8, 899 (2014).
 6

8 34
G. Wang, A. Chernikov, M. M. Glazov, T. F. P. Xie, Z. Liang, T. Jia, D. Li, Y. Chen,
Heinz, X. Marie, T. Amand, and B. Urbaszek, P. Chang, H. Zhang, and W. Wang,
Rev. Mod. Phys. 90, 021001 (2018). Phys. Rev. B 104, 125446 (2021).
9 35
C. Chen, M. L. Juan, Y. Li, G. Maes, G. Borghs, P. V. I. A. M. Al-Ani, K. As’ham, L. Huang,
Dorpe, and R. Quidant, Nano Lett. 12, 125 (2012). A. E. Miroshnichenko, and H. T. Hattori,
10
T. T. Tran, D. Wang, Z.-Q. Xu, A. Yang, M. Toth, T. W. Laser Photonics Rev. 15, 2100240 (2021).
36
Odom, and I. Aharonovich, Nano Lett. 17, 2634 (2017). J. Lin, M. Qiu, X. Zhang, H. Guo, Q. Cai, S. Xiao, Q. He,
11
A. I. Kuznetsov, A. E. Miroshnichenko, M. L. and L. Zhou, Light Sci. Appl. 9, 158 (2020).
37
Brongersma, Y. S. Kivshar, and B. Luk’yanchuk, A. Raman and S. Fan,
Science 354, aag2472 (2016). Phys. Rev. Lett. 104, 087401 (2010).
12 38
K. Koshelev, G. Favraud, A. Bogdanov, Y. Kivshar, and B. Xi, H. Xu, S. Xiao, and L. Zhou,
A. Fratalocchi, Nanophotonics 8, 725 (2019). Phys. Rev. B 83, 165115 (2011).
13 39
K. Koshelev and Y. Kivshar, S. Fan, W. Suh, and J. D. Joannopoulos,
ACS Photonics 8, 102 (2020). J. Opt. Soc. Am. A 20, 569 (2003).
14 40
C. W. Hsu, B. Zhen, J. Lee, S.-L. Chua, S. G. L. Xu, K. Z. Kamali, L. Huang, M. Rahmani,
Johnson, J. D. Joannopoulos, and M. Soljačić, A. Smirnov, R. Camacho-Morales, Y. Ma, G. Zhang,
Nature 499, 188 (2013). M. Woolley, D. Neshev, and A. E. Miroshnichenko,
15
K. Koshelev, S. Lepeshov, M. Liu, A. Bogdanov, and Adv. Sci. 6, 1802119 (2019).
41
Y. Kivshar, Phys. Rev. Lett. 121, 193903 (2018). X. Wang, J. Duan, W. Chen, C. Zhou, T. Liu, and S. Xiao,
16
S. Li, C. Zhou, T. Liu, and S. Xiao, Phys. Rev. B 102, 155432 (2020).
42
Phys. Rev. A 100, 063803 (2019). S. Sun, Z. Zhou, C. Zhang, Y. Gao, Z. Duan, S. Xiao, and
17
L. Huang, L. Xu, M. Rahmani, D. Neshev, and A. E. Q. Song, ACS Nano 11, 4445 (2017).
43
Miroshnichenko, Adv. Photonics 3, 016004 (2021). B. Yang, W. Liu, Z. Li, H. Cheng, D.-Y. Choi, S. Chen,
18
Y. Kang, S. Najmaei, Z. Liu, Y. Bao, Y. Wang, X. Zhu, and J. Tian, Nano Lett. 19, 4221 (2019).
44
N. J. Halas, P. Nordlander, P. M. Ajayan, J. Lou, and F. Todisco, R. Malureanu, C. Wolff, P. A. D. Gonçalves,
Z. Fang, Adv. Mater. 26, 6467 (2014). A. S. Roberts, N. A. Mortensen, and C. Tserkezis,
19
S. Zu, B. Li, Y. Gong, Z. Li, P. M. Ajayan, and Z. Fang, Nanophotonics 9, 803 (2020).
45
Adv. Opt. Mater. 4, 1463 (2016). Y. Li, A. Chernikov, X. Zhang, A. Rigosi, H. M. Hill, A. M.
20
J. Sun, H. Hu, D. Zheng, D. Zhang, Q. Deng, S. Zhang, van der Zande, D. A. Chenet, E.-M. Shih, J. Hone, and
and H. Xu, ACS Nano 12, 10393 (2018). T. F. Heinz, Phys. Rev. B 90, 205422 (2014).
21 46
S. Zhang, H. Zhang, T. Xu, W. Wang, Y. Zhu, S. Xiao, T. Liu, X. Wang, X. Liu, and C. Zhou,
D. Li, Z. Zhang, J. Yi, and W. Wang, Phys. Rev. B 102, 085410 (2020).
47
Phys. Rev. B 97, 235401 (2018). C. Zhou, X. Qu, S. Xiao, and M. Fan,
22
K. L. Koshelev, S. K. Sychev, Z. F. Sadrieva, A. A. Bog- Phys. Rev. Appl 14, 044009 (2020).
48
danov, and I. V. Iorsh, Phys. Rev. B 98, 161113 (2018). J. Zhang, K. F. MacDonald, and N. I. Zheludev,
23
E. S. H. Kang, S. Chen, S. Sardar, D. Tordera, N. Ar- Opt. Express 21, 26721 (2013).
49
makavicius, V. Darakchieva, T. Shegai, and M. P. Jonsson, N. Kongsuwan, A. Demetriadou, R. Chikkaraddy, F. Benz,
ACS Photonics 5, 4046 (2018). V. A. Turek, U. F. Keyser, J. J. Baumberg, and O. Hess,
24
W. Du, J. Zhao, W. Zhao, S. Zhang, H. Xu, and Q. Xiong, ACS Photonics 5, 186 (2017).
50
ACS Photonics 6, 2832 (2019). X. Liu, T. Galfsky, Z. Sun, F. Xia, E. chen Lin,
25
S. Wang, Q. Le-Van, F. Vaianella, B. Maes, S. E. Y.-H. Lee, S. Kéna-Cohen, and V. M. Menon,
Barker, R. H. Godiksen, A. G. Curto, and J. G. Rivas, Nat. Photonics 9, 30 (2014).
51
ACS Photonics 6, 286 (2019). X. Han, K. Wang, X. Xing, M. Wang, and P. Lu,
26
H. Li, M. Qin, Y. Ren, and J. Hu, ACS Photonics 5, 3970 (2018).
52
Opt. Express 27, 22951 (2019). V. Kravtsov, E. Khestanova, F. A. Benimetskiy,
27
P. Xie, Z. Liang, Z. Li, W. Wang, W. Wang, T. Ivanova, A. K. Samusev, I. S. Sinev, D. Pidgayko,
T. Xu, X. Kuang, L. Qing, D. Li, and J. Yi, A. M. Mozharov, I. S. Mukhin, M. S. Lozhkin, Y. V.
Phys. Rev. B 101, 045403 (2020). Kapitonov, A. S. Brichkin, V. D. Kulakovskii, I. A.
28
S. Cao, H. Dong, J. He, E. Forsberg, Y. Jin, and S. He, Shelykh, A. I. Tartakovskii, P. M. Walker, M. S.
J. Phys. Chem. Lett. 11, 4631 (2020). Skolnick, D. N. Krizhanovskii, and I. V. Iorsh,
29
Y. Chen, S. Miao, T. Wang, D. Zhong, A. Saxena, Light Sci. Appl. 9, 56 (2020).
53
C. Chow, J. Whitehead, D. Gerace, X. Xu, S.-F. Shi, and J. R. Piper and S. Fan, ACS Photonics 1, 347 (2014).
54
A. Majumdar, Nano Lett. 20, 5292 (2020). L. Huang, G. Li, A. Gurarslan, Y. Yu, R. Kirste, W. Guo,
30
P. Xie, D. Li, Y. Chen, P. Chang, H. Zhang, J. Yi, and J. Zhao, R. Collazo, Z. Sitar, G. N. Parsons, M. Kudenov,
W. Wang, Phys. Rev. B 102, 115430 (2020). and L. Cao, ACS Nano 10, 7493 (2016).
31 55
L. Zhao, Q. Shang, M. Li, Y. Liang, C. Li, and Q. Zhang, J. Nie, J. Yu, W. Liu, T. Yu, and P. Gao,
Nano Res. 14, 1937 (2020). Opt. Express 28, 38592 (2020).
32 56
K. As’ham, I. Al-Ani, L. Huang, A. E. Miroshnichenko, F. Wu, D. Liu, and S. Xiao,
and H. T. Hattori, ACS Photonics 8, 489 (2021). Opt. Express 29, 41975 (2021).
33
M. Qin, S. Xiao, W. Liu, M. Ouyang, T. Yu, T. Wang,
and Q. Liao, Opt. Express 29, 18026 (2021).