Xuchen
Nie
*a,
Xiaoyue
Wu
a,
Yang
Wang
a,
Siyuan
Ban
a,
Zhihao
Lei
b,
Jiabao
Yi
b,
Ying
Liu
*c and
Yanpeng
Liu
*a
aKey Laboratory for Intelligent Nano Materials and Devices of Ministry of Education, State Key Laboratory of Mechanics and Control of Mechanical Structures, and Institute for Frontier Science, Nanjing University of Aeronautics and Astronautics, Nanjing, 210016, China. E-mail: xcnie@nuaa.edu.cn; chmliuyp@nuaa.edu.cn
bGlobal Innovative Centre for Advanced Nanomaterials, College of Engineering, Science and Environment, The University of Newcastle, NSW 2308, Australia
cCollege of Jincheng, Nanjing University of Aeronautics and Astronautics, Nanjing 211156, China. E-mail: lily_ying2012@163.com
First published on 12th November 2022
Surface acoustic wave (SAW)–matter interaction provides a fascinating key for inducing and manipulating novel phenomena and functionalities in two-dimensional (2D) materials. The dynamic strain field and piezo-electric field associated with propagating SAWs determine the coherent manipulation and transduction between 2D excitons and phonons. Over the past decade, many intriguing acoustic-induced effects, including the acousto-electric effect, acousto-galvanic effect, acoustic Stark effect, acoustic Hall effect and acoustic exciton transport, have been reported experimentally. However, many more phenomena, such as the valley acousto-electric effect, valley acousto-electric Hall effect and acoustic spin Hall effect, were only theoretically proposed, the experimental verification of which are yet to be achieved. In this minireview, we attempt to overview the recent breakthrough of SAW-induced phenomena covering acoustic charge transport, acoustic exciton transport and modulation, and coherent acoustic phonons. Perspectives on the opportunities of the proposed SAW-induced phenomena, as well as open experimental challenges, are also discussed, attempting to offer some guidelines for experimentalists and theorists to explore the desired exotic properties and boost practical applications of 2D materials.
2D materials, owing to their atomic thickness and extremely large surface-to-volume ratio, allow external stimuli (electric fields and SAWs) to modulate their inherent optical, electrical, and opto-electronic properties.37–44 In 2D limit, dielectric screening is reduced while Coulomb interactions and excitonic emissions (exciton, trion and biexciton) are extremely enhanced.45 However, the atomic thickness of 2D materials signifies indispensable supporting substrates for most fundamental research and applications. Therefore, the coupling between 2D materials and underlying substrates has blossomed into one promising key for controlling 2D material properties. In this regard, by placing 2D materials onto piezo-electric substrates, the as-generated SAWs under electrical excitations as well as the corresponding dynamic strains and piezo-electric fields concentrated at substrate surfaces were proven to modulate the lattices and electronic structures of 2D materials in an elegant manner. Moreover, for certain 2D materials with trigonal prismatic structures, such as MoS2 and WS2 monolayers, SAWs could break the inversion symmetry and give impetus to fantastic piezo-electricity and strain modulations.46–48
Herein, we review the research progress of SAW-induced emergent phenomena (as shown in Fig. 1) in 2D materials, such as semi-metallic graphene and semiconducting transition metal dichalcogenides (TMDCs), mainly excited by interdigital transducers (IDTs) on piezo-electric substrates. This minireview is organized as follows: first, we brief the basic coupling feature between SAWs and 2D materials; then, we classify our discussions into Type-I and Type-II band edge modulations by dynamic strains and piezo-electric fields associated with SAW propagation. The confronted challenges and feasible future directions are also provided. We hope this timing review will offer intuitive understanding of the SAW-induced emergent phenomena in 2D materials and also provide possible opportunities for both new comers and experiential experts.
To obtain desirable SAWs, electrical excitations based on the inverse piezo-electric effect50–54 and optical excitation in light of elastic stresses upon laser radiation55–57 are two prevalent methods. For the former one, SAWs are mainly generated, detected, and controlled by employing IDTs. IDTs are periodic arrays of metallic finger electrodes with a pitch of half the SAW wavelength (λSAW). Two interlocking electrodes are deposited on piezo-electric substrates and connected to a frequency generator.58 Piezo-electric crystals with dipoles re-oriented along certain directions under external electric fields, such as quartz, lithium niobate (LiNbO3) and lithium tantalate (LiTaO3), generate directional propagations of mechanical deformation waves along piezo-electric surfaces. As shown in the inset in Fig. 2, acoustic actuation is launched by applying a high frequency input signal between transmitting IDTs to achieve SAWs with a targeted frequency (f = νSAW/λSAW, where νSAW and λSAW represent the velocity and wavelength of SAWs), complying with the inverse piezo-electric effect. Because of the mechano-electrical coupling of piezo-electric substrates, the IDTs at the receiving end produce electrical signals, referred to as acoustic sensing. By designing the spacing and geometry of electrodes, appropriate SAWs could be selectively emitted toward distinct applications.58–60 In particular, the dispersive delay line (chirped) configuration with varying spacings could produce altered frequencies between electrodes that range from 20 MHz to 125 MHz bandwidths.61 Moreover, the periodic waveforms of high SAW harmonics with arbitrary shapes might also be manifested by changing the unit cell patterns.62 For the majority of electrical SAW excitations, the dominant working principle is based on the inverse piezo-electric effect. In a stress-free planar surface of a semi-infinite isotropic half-space, SAW modes could be expressed by solving the elastic wave equation for the acoustic displacement field
(
, t) that decays towards zero as it penetrates into the bulk counterpart63
![]() | (1) |
-axis, the solutions of the elastic equation must also satisfy the constitutive relationshipsT = cS − d![]() |
= dS + ε![]() | (2) |
| Tiz(z = 0) = 0, (i = x,y,z) |
| Dz(z = 0+) = Dz(z = 0−) | (3) |
is the electric displacement field and
is the piezo-electric field (expressed in terms of the piezo-electric potential φSaw). To our limited knowledge, the drawback of electrical excitation methods is the complexity of matching large SAW amplitudes with small quantum objects. Moreover, broadband SAWs are strictly prohibited in IDTs and the current piezo-electric materials are unsatisfying. The key properties of appropriate piezo-electric materials include the electro-mechanical coupling factor (k2) that determines the piezo-electric coupling efficiency, the coefficient of thermal expansion, the acoustic velocity (νSAW), compatibility, and finally the expense of raw materials.2Table 1 summarizes the popular piezo-electric materials and their corresponding parameters. For certain bulk piezo-electric materials, orientational cutting may generate particular types of SAWs and affect the acoustic velocity. For example, 128° and 41° YX cutting of LiNbO3 were found to generate Rayleigh and shear-horizontal waves, respectively. The acoustic velocity of the LiNbO3 crystal was increased to 3992 m s−1 (128° YX cutting one) from 3488 m s−1 (pristine, no orientational cut).2 Thin-film piezo-electric materials, including aluminum nitride (AlN), zinc oxide (ZnO), and piezo-ceramics, have also attracted broad attention in lab-on-chip and acousto-fluidics devices due to their flexibility and selective deposition compatibility.64–67 Among them, AlN films were reported to exhibit the highest acoustic velocity (5800 m s−1) and high temperature resistance,65 while piezo-ceramics exhibit the largest piezo-electric constant (d33 = 110 pC/N for PZT, lead zirconate titanate) and electro-mechanical coupling.67
![]() | ||
Fig. 2 Acousto-electric effect. (center) Device schematics of SAW excitation and detection with a basic delay line electrode configuration consisting of two IDTs on piezo-electric substrates. The left IDT generates a surface Rayleigh wave propagating along the axis that is read out by right IDTs. The insets show the electrical wiring configurations for 4-point (left) and 2-point (right) measurements of the acoustic charge transport. (a and b) Measured SAW phase and AE current vs. gate voltages (Vgate) in graphene/LiNbO3 devices. Inset: Schematic diagram of the device. (c) AE current as a function of the SAW intensity (at 32 MHz) for three different Vgate. Reproduced from ref. 97 with permission from Tsinghua University Press, copyright 2016 (Open access). (d) AE voltage vs. radio frequency (RF) power plots of the hybrid monolayer MoS2/LiNbO3 device via a 4-point open-circuit (ISD = 0) configuration. For forward and backward SAW propagation directions, the expected linear dependences are well reproduced. The signs of AE currents and voltages correspond to the n-type conductivity of the film. (e) SAW transmissions and propagations. The frequency band of SAW transmission between IDTs plotted as the scattering parameter S21 (top panel). The AE current as a function of the applied radio frequency for different RF power levels in a 2-point short-circuit (VSD = 0) configuration (bottom panel). Reproduced from ref. 98 with permission from Springer Nature, copyright 2015 (Open access). (f) Schematic layout of the IDTs and BP/MoS2 p–n heterojunction. (g) Photocurrent (Iph) variations at different RF powers under a light illumination of 100 mW. As the applied RF power increases to 21 dBm, Iph increases sharply to 900 nA, overwhelming the value in the absence of SAWs (∼280 pA). (h) Photo-responsivity comparison with and without SAW excitations at different laser powers. Reproduced from ref. 99 with permission from Royal Society of Chemistry, copyright 2018. (i and j) Temperature and magnetic-field dependent resistances of the NbSe2 device under different excitation powers at 1.6 K. (k) Schematics showing the nucleation of soliton and antisoliton pairs. One possible scenario for reproducing the negative resistance is irradiating SAWs to generate soliton–antisoliton pairs in the CDW phase, which forms local capacitances in the superconducting domains. When SAWs irradiate the CDW states of NbSe2 (below 33 K), all selenium atoms would be shifted from the commensurate position, and thus, the CDW phase φ is modulated over the SAW wavelength (top panel). This displacement increases the electrostatic energy and elastic energy. Thus, in the superconducting state (below Tc = 7.2 K), it is energetically more favorable to nucleate soliton–antisoliton pairs and induce a 2π phase difference between the soliton and antisoliton pairs (bottom panel). Reproduced from ref. 100 with permission from American Association for the Advancement of Science, copyright 2020. | ||
| Piezo-electric materials | Orientation/cut | Piezo-electric coefficient (d33, pC/N) | Acoustic velocity (m s−1) | Density (g cm−3) | Thermal expansion coefficient (×10−6/°C) | Elastic modulus (GPa) | Electro-mechanical coupling coefficient (k2, %) |
|---|---|---|---|---|---|---|---|
| Quartz | X axis | 2.3 (d11) | 3159 | 2.65 | — | 97.2 | 0.14 |
| Quartz | ST-cut | 2.3 (d11) | 3159 | 2.65 | — | 71.7 | 0.03 |
| LiNbO3 | YZ-cut | 6 | 3488 | 4.65 | — | 202 (C11) | 0.045 |
| LiNbO3 | 128°YX -cut | 12 | 3992 | 4.65 | — | 202 (C11) | 5.3 |
| LiTaO3 | YZ-cut | 8 | 3230 | 7.46 | — | 233 (C11) | 0.66 |
| LiTaO3 | X-cut | 5.7 | 3290 | 7.46 | — | 233 (C11) | 0.75 |
| BaTiO3 | — | 160 | 4392 | 5.70 | — | 110 | 0.34 |
| PZT | — | 110 | 3900 | 7.50 | — | 60 | 20–35 |
| ZnO | — | 12 | 2720 | 5.61 | 6.50 | 140 | 1.5 |
| GaAs | — | 3.4 | 4730 | 5.32 | 5.73 | 86 | 0.07 |
| GaN | — | 4.5 | 4130 | 6.15 | 3.17 | 320 | 0.13 |
| AlN | — | 4.5, 6.4 | 5800 | 3.30 | 5.20 | 300–350 | 3.1–8 |
With reference to electrical excitations, the optical excitations are reported to be capable to obtain SAWs with spectral widths up to several GHz (known as acoustic phonons).55,68 In pump–probe configurations, typically, the picosecond/femtosecond laser is split into pump and probe pulses, where the pump pulse generates ultrafast ultrasonic waves into target samples via photo-thermal and thermo-mechanical couplings, and then the probe pulse reflected from the sample surface is used to measure the pump-induced ultrasonic waves in the GHz to THz range by interferometric and reflectometric methods.69,70 For the interferometric method, the Michelson, Mach–Zehnder, and Sagnac interferometric techniques are preferred to measure the absolute optical phase of the reflected probe pulse that is sensitive to the surface displacements.71–74 Therefore, outstanding sensitivity and spatial resolution to very tiny acoustic vibrations in the ultra- and super-high-frequency ranges could be expected.74,75 However, the implementation of interferometry in the picosecond/femtosecond time scale demands precise stabilization systems, and the phase calibration becomes complicated. In contrast, the reflectometric method measures the small intensity variation of probe reflection due to the photo-elastic coupling of material reflectivity with ultrasonic waves. For ensuring the simplicity in implementations, lock-in amplifying by using knife-edge76 and beam distortion detection77 techniques that precisely capture the surface displacements are the mainstreams. By introducing local heating and thermo-elastic expansion on 2D surfaces via focused ultra-short laser pulses, broadband SAW pulses might be excited within any absorbing material in the ultrasound region in theory. The mere limitation of the SAW amplitude relies on the damage threshold of 2D materials. In a nutshell, the optical excitation method paves the way for systematically investigating the linear and nonlinear elastic properties of 2D materials and their heterostructures in the frame of thermo-elastic and ablative or explosive evaporation mechanisms, respectively.
![]() | (4) |
is the phonon pressure, I is the SAW intensity, and Γ is the attenuation coefficient given by82,83![]() | (5) |
This is a non-monotonic function of the diagonal component of a conductivity tensor σ2D and reaches a maximum at the characteristic conductivity σM. During last few years, the AE effect in 2D materials has been extensively investigated toward applications as metrology and quantum information processing.104 For instance, the SAW excitations of graphene generated a direct AE current in graphene/piezo-electric substrates.63 The AE current (Fig. 2a–c) was found to increase linearly with SAW frequencies and powers82,83 and could be reversed and switched off by a gate voltage.97 The flow of induced currents and the accumulation of longitudinal voltages in graphene were further verified under different experimental conditions, including varying temperatures, carrier concentrations, piezo-electric substrates and wiring schemes.81–86 The coupling of piezo-electric fields with electron gas modifies the propagation velocity, and even amplify the SAW intensity within a graphene plane.97,105–107 Through illuminating a hybrid graphene/LiNbO3 structure, Tynyshtykbaev et al. pioneerly reported the SAW-induced photo-acoustoelectric amplification in 2021. The appearance of an additional current and reduced collisional scattering of photo-induced electrons under the piezo-electric field proved the origin of the SAW-induced photo-acoustoelectric effect.107 In addition, the periodic electric and strain fields of SAWs were also adopted to manifest the graphene Dirac cone108 and even couple lights into plasmons.109
In afore-mentioned studies, the majority of 2D materials interplaying with SAWs were mechanically exfoliated from mother crystals and the lateral size was usually limited to below few tens of millimeters. In addition, several unavoidable transfer processes might have introduced wrinkles and tear 2D materials into smaller pieces, adding to difficulty in massive productions. In this concern, epitaxial growth of 2D materials onto SAW substrates (quartz and LiNbO3) was attempted. Preciado et al.98 successfully developed a millimeter-scale MoS2/LiNbO3 structure by chemical vapor deposition (CVD) methods, followed by lithographically defining a field-effect transistor (FET) structure on the top. The acousto-electric activity of the LiNbO3 substrate permits concomitant probing and manipulating the electronic and optical properties of 2D materials in a contact-free manner. Fig. 2d and e show the strong acousto-electric effect and enhanced photo-conductance of the epitaxial monolayer MoS2 on the LiNbO3 substrate.
Stimulated by the success of single 2D materials, Zheng et al.99 proposed a p–n diode by fabricating a BP/MoS2 van der Waals (vdWs) heterojunction onto SAW platforms. Under the pronounced piezo-electric fields, the p–n heterojunction diode achieved an ultra-high photo-responsivity of 2.17 A W−1 at a wavelength of 582 nm (Fig. 2f–h), nearly 103 times higher than the value achieved without SAWs. Departing from the monolayer MoS2/Au nanoparticles, a Love-mode SAWs aptasensor with dummy fingers was developed and exhibited sensitive and rapid detections of alphafetoprotein in serum in the range of 0.01–100 ng mL−1 with the lowest detection limit of 4.79 pg mL−1.110 The demonstration of sensitive visible-light and protein detection unravels the feasibility of implementing SAWs into next-generation photodiodes and sensors.
Owing to the concurrence of broken inversion symmetry and strong spin–orbit coupling in TMDC monolayers, the electrons (or excitons) at two valleys (K or K′) in the momentum space show opposite spin states with equal energy. Hence, the carriers from extraordinary K (K′) valley states couple exclusively with right (left) circularly polarized light in accordance with the valley-dependent optical selection rule. The coherent control of valley pseudospin by polarized lights provides a nontrivial degree of freedom to manipulate carriers and encode information. Because of the nonzero symmetrical Berry curvature, the in-plane electric field will give impetus to Hall currents with the valley-index dependent sign, defined as the valley Hall effect.37,111 In addition, by placing non-centrosymmetric TMDC monolayers onto piezo-electric substrates, the valley acousto-electric effect and the valley acousto-electric Hall effect were predicted using SAWs. The essence of the valley acousto-electric effect lies in the emergence of a drag electric current (and a spin current) and the contribution of the as-generated current was assigned into three parts: one is the conventional diffusion current that is independent of the valley index and proportional to the acoustic wave vector; the other is the Hall-like warping current due to the trigonal warping of electron dispersion; the third is the acousto-electric valley Hall current caused by the Berry phase acquired by Bloch electrons traveling along the 2D crystal.94,95
In the community of strongly correlated electronic 2D systems, SAWs were attempted to modulate the interplay between charge-density wave (CDW) and superconductivity. In superconducting NbSe2 thin films on the LiNbO3 substrate, Yokoi et al.100 reported the SAW-induced negative resistance (a voltage decrease along the opposite direction of the current flow at zero bias) and found that the amplitude of negative resistance increased by elevating the SAW power and lowering the temperature (Fig. 2i–k). The observation of negative resistance states might become essential in verifying SAW-driven Floquet engineering in superconducting states. In heavy-fermion chiral superconductors, the coupling of SAWs with clapping modes, which are relative amplitude and phase modes between the two chiral components of Cooper pairs, also generates an anomalous transverse electric current in addition to the alternating AE current.85 This anomalous AE effect in chiral superconductors is supposed to be a universal feature and therefore, more research investments may expedite fantastic SAW-induced phenomena.
![]() | ||
| Fig. 3 Acousto-galvanic effect. (a) Device configuration of acoustic transport in graphene. A microwave frequency is applied to the IDTs. The monolayer graphene on hybrid piezo-electric substrates was patterned into Hall bar geometry and the acoustic current (ISAW) was measured in the 2-point short-circuit configuration. (b) Plots of ISAW evolution as a function of the IDT frequency, power, and back gate voltage (VBG). (c) Acoustically induced voltages Vxx as a function of the IDT power for the fundamental (f0), first (f1) resonance and three different carrier concentrations at B = 0 T (i.e. the electron regime, the charge neutrality point (CNP), and the hole regime controlled by VBG). The sign of Vxx reflects the dominating carrier type in the graphene layer and is negative (positive) when the Fermi energy resides in the conduction (valence) bands. (d and e) The longitudinal voltage Vxx and transverse voltage component Vxyversus the IDT power, resonances and gate voltages at 4.2 K. The acoustically induced giant synthetic transverse Hall voltage Vxy is positive in the electron regime, approaches zero around the CNP, and then changes the sign once entering the hole regime. The dashed line marks the charge neutral point. Reproduced from ref. 87 with permission from American Physical Society, copyright 2022. | ||
Type-II spatial carrier separation could be induced by the piezo-electric potential along the SAW propagation direction, where the piezo-electric potential ionizes the optically excited excitons and transports the as-generated electrons (holes) within the potential maxima (minima). In this way, the acoustic-induced photoluminescence (PL) quenching arises from exciton ionization (i) followed by the spatial separation of the free electrons and holes (ii) and transport out of the detection area (iii). In addition, the SAWs can increase the radiative recombination lifetime and thus may favor non-radiative recombination channels for PL quenching. The exciton ionization rate depends on the field amplitude and conversely, the travelling piezo-electric field associated with Rayleigh SAW managed the reversible manipulation of excitonic PL in 2–4 layered MoS2 nanoflakes. It is shown that the interplay between photo-generated carriers, excitons, and the SAW piezo-electric field leads to strong dependence of the PL intensity on the SAW amplitude and excitation intensity.121 Subsequently, at moderate SAW powers, researchers from the same group118 observed a strong PL quenching followed by an energy blue-shift of the neutral and negatively charged exciton (trion) emission bands (Fig. 4a–d). The dissociation and spatial separation of excitons were also observed because of the acoustically generated electric field arising from the intrinsic piezo-electricity of odd layers of MoS2 flakes.122 This intrinsic piezo-electricity may be quite promising in improving the electron–hole dissociation efficiency in photovoltaic cells. In 2021, Sheng et al.123 further demonstrated that the exciton characteristics strongly depend on the layer numbers of MoS2 sheets. Under acoustic excitations, the PL quenching in odd MoS2 layers is remarkably larger than that in even layers without piezo-electricity. Moreover, the PL peak position of odd MoS2 layers is blue-shifted with an increment of acoustic powers but remains unchanged in even MoS2 layers.
![]() | ||
| Fig. 4 Acoustic exciton modulations. (a) PL mapping of monolayer MoS2 flakes before and after SAWs. The acoustic exciton dissociation and transport in 2D semiconductor materials exposed to SAWs, directly monitored through spatially and time-resolved PL microscopy. (b) PL spectra evolution of monolayer MoS2 flakes by changing SAW input powers. At input powers ≤150 mW, strong quenching can be observed together with an apparent blue-shift of the PL peak due to the ionization of trions into excitons by electric fields, triggered as a consequence of the inherent piezo-electricity by mechanical vibrations using SAWs. At input powers ≥0.25 W, heat generation becomes significant and shifts the PL peak toward higher wavelengths (red-shift). (c and d) PL spectra of bilayer and trilayer MoS2 flakes with and without SAWs. Reproduced from ref. 118 with permission from American Chemical Society, copyright 2016. (e) Illustration showing the conduction and valence band potential fluctuations along the carrier transport path induced by piezo-electric fields. (f) PL centroid map of the WSe2 monolayer at −60 dBm and 20 dBm power levels. (g) Histogram of the PL centroid extracted from the PL mapping (f). Reproduced from ref. 89 with permission from American Chemical Society, copyright 2021. (h) PL spectra of the monolayer MoSe2 for a series of in-plane SAW fields (Ex) ranging from 0 to 50 kV cm−1. (i) The negatively charged trions (X−) and neutral exciton (X0) emissions red-shift ΔE as a function of Ex. The dashed cyan and orange lines are the Stark shift contributions ΔEStark to ΔE extracted from fits. Reproduced from ref. 90 with permission from American Chemical Society, copyright 2021. | ||
Fig. 4e–g show the interaction of excitons in the monolayer WSe2 with the SAW piezo-electric field (type-II bandgap modulation) at room temperature investigated by PL spectroscopy. As can be seen, intense PL quenching is observed with increasing SAW powers because of the dissociation of excitons and trions into free carriers by the in-plane electric field. The PL centroid map and the corresponding histogram clearly show red-shift under SAW excitation, which is attributed to the acoustic Stark effect. By utilizing photo-generated free carriers, the author further demonstrated that the exciton-piezo-electric field interaction and population distribution between neutral excitons and trions could be optically manipulated by the screening field.89 Similarly, Fig. 4h–i show the successfully manipulating excitonic emission energies (i.e. the acoustic Stark effect) and dissociate excitonic states in the monolayer MoSe2 by the in-plane piezo-electric field at 5 K. The high dielectric constant of the LiNbO3 substrate strongly reduces the neutral exciton binding energy, leading to efficiently quench (above 90%) and red-shift of the excitonic emission energies.90 Additionally, the strong coupling between excitons in 2D semiconductors and surface plasmons from 2D materials was predicted to produce SAW-accessible exciton–plasmon polaritons. The strain fields of SAWs create optical diffraction gratings that provide the momentums matched with the surface plasmon, canceling out the piezo-electric field that could dissociate excitons. This is exemplified for the MoS2 monolayer and black phosphorus (BP) mono- and few-layer placed on the top of the thin silver/LiNbO3 substrate.124
Defect centers in 2D materials possess atomic-like optical emission and could couple efficiently with SAWs and related strain and piezo-electric fields in the frequency range of hundreds of MHz. Similar to external RF fields, SAWs can modulate the emission energies and induce resonant transitions between their spin sublevels. Coherent interactions between SAWs and defect centers have been demonstrated for single negatively charged nitrogen-vacancy17,35 and silicon-vacancy34,125 centers in diamond, and neutral di-vacancy centers in silicon carbide (SiC).20 In terms of 2D materials, Iikawa et al.126 investigated the dynamic modulation of optically active defect centers in hexagonal boron nitride (h-BN) layers using SAWs. The dynamic strain field of SAWs modulates the emission lines, resulting in intensity variations as large as 50% and oscillations of the emission energy with an amplitude of ∼1 meV.
The spatial–temporal dynamics of exciton generation, annihilation and spatial diffusion could be separately investigated due to their different time scales. Assuming that none in-plane electric field is applied and the many-body interactions between exciton–trion and exciton–phonon could be ignored, the spatial–temporal dynamics of exciton can be written as127
![]() | (6) |
is the position vector (x and y are in-plane coordinates), Δ = ∂2/∂x2 + ∂2/∂y2 is the Laplace operator, τ is the exciton recombination lifetime, Deff is the effective diffusion coefficient of the exciton, and β is the exciton–exciton annihilation rate, respectively. The second and third terms represent the diffusion and Auger recombination, respectively. For exciton diffusion in the low-density regime (i.e. linear diffusion) that the Auger process (∼βN2) was ignored; eqn (6) can be simplified as![]() | (7) |
The initial exciton density excited by laser pulses is assumed as N0 and the analytical solution is expressed as
, where
is the spatial distribution width of the exciton density and w0 is the width of excitation beams. In the high-density regime (i.e. nonlinear diffusion), the Auger process is remarkably faster than exciton diffusion. By ignoring the spatial diffusion of excitons, thus eqn (6) can be simplified as
![]() | (8) |
. By using the solution of eqn (7) as the initial spatial–temporal distribution of excitons, the nonlinear diffusion could be approximately solved with the effective diffusion coefficient Deff = D + βN0w02/16, where D is the intrinsic diffusion coefficient.130 In addition to the exciton–exciton annihilation in the nonlinear diffusion regime, the repulsive exciton interactions that composed the dipolar repulsion and exchange interaction may also significantly improve Deff of the exciton. The exchange interaction decreases with increasing electron–hole separation and becomes negative when the vertical separation is larger than the Bohr radius. For IXs in TMDC heterostructures, the electron–hole separation approximates the Bohr radius of IXs (∼1 nm), thus the exchange interaction is negligible. After taking the repulsive exciton interaction into account, the drift-diffusion equation becomes capable to quantitatively describe the spatial–temporal evolution of the IX density![]() | (9) |
For excitonic devices, an outstanding exciton transport distance with a controlled directionality is desired to transmit information over sizable distances. However, a moderate in-plane electric field may dissociate excitons due to its small binding energy, extremely limiting the practical application of excitonic devices. Generally, excitons in quasi-0D systems (molecular and quantum dots) have low diffusion lengths (1–50 nm)151 at room temperature, as the diffusion mainly occurs by site hopping. One-dimensional (1D) single-wall carbon nanotubes have larger exciton diffusion lengths (∼300 nm), depending on the chirality and dielectric environment.144,152,153 In bulk semiconductors, strong dielectric screening generally leads to low exciton binding energy, so excitons exist (maximum diffusion length ∼10–1000 μm) only at cryogenic temperatures,120,131,132 unsatisfying for practical applications.
Owing to the reduced dielectric screenings, enhanced Coulomb interactions, and relatively large effective masses of charge carriers, excitons in 2D TMDC monolayers have a Bohr radius of ∼1 nm and a binding energy of ∼500 meV with strong resonances even at room temperature, making them promising for a plethora of opto-electronic and quantum applications.127 Furthermore, even larger exciton diffusion lengths can be achieved in TMDC-based devices, either by separating electrons and holes in different layers in type-II vdWs heterostructures (where the conduction band minimum and valence band maximum are located in two different layers),154–159 or by using inhomogeneous strain profiles to funnel excitons (i.e. the exciton funneling effect)133–137 that largely reduce exciton–phonon scattering.160,161 The separation between the electron and hole layers allows controlling the overlap of electron and hole wave functions and achieving long IX lifetimes, orders of magnitude longer than lifetimes of spatially direct excitons (DXs, also referenced as intralayer excitons).162 In other words, IXs travel longer distances before recombination. In addition, for MoS2/WS2 and WSe2/WS2 heterostructures with the type-II band alignment, photoexcited DXs turned into IXs on ultrafast timescales of less than 50 fs, promising toward excitonic devices operating at room temperature or even higher temperatures.163,164 Another appealing strategy to control and enhance exciton transport could be generating a time-varying (dynamic) inhomogeneous strain profile by SAWs.134 The spatial modulation of both the conduction and valence bands with opposite phases (type-I bandgap modulation) creates periodic spatiotemporal traps that behave as an efficient conveyor for excitons. The propagating SAWs from the piezo-electric substrate will generate a near-field piezo-electric field with an out-of-plane field (Ez) amplitude up to 107 V m−1 at an acoustic power density of 1 mW μm−1. As such, the Ez component will periodically modulate the IX energy in space and time, creating a dynamic trapping potential for IXs in the extrema of Ez. As the SAWs propagate, the IXs will drift along the time-varying gradient of Ez and thus be carried by the SAWs like surfing on water waves, traveling a longer distance before recombination. In addition, SAWs might also permit us to generate strain along multiple axes and tailor the strain fields. For example, a biaxial strain pattern could be generated by either using orthogonal SAWs or selecting crystal cuts and propagation directions.91
To control the directional transport of the charge neutral exciton flux, spatial tunings of exciton potential by mechanical strain160,165 or an electric field139–142,166 were realized and are summarized in Table 2. Travelling SAWs can dynamically utilize both strain fields and piezo-electric fields to manipulate and transport excitons in 2D semiconducting materials. SAWs enable the creation of tunable strain fields with micrometer-sized dimensions have been employed to effectively transport IXs with a controlled directionality in III–V semiconductor QWs119,120,131,132,167 and 2D vdWs heterostructures.91,138 Several hundred micrometers of the IXs transport distance in GaAs QWs subjected to a transverse electric field has been achieved at 12 K.131 However, the low exciton binding energy of only few meV in GaAs prohibits the operation at higher temperatures. The employment of SAWs to manipulate the exciton transport of monolayer TMDCs and heterostructures may be one future direction. Nowadays, it shows that SAWs can direct the flux of excitons in the TMDC monolayer over micrometer distances, even at room temperature. Fig. 5d–f show the room-temperature controlled and directed exciton transport in the monolayer WSe2 by adopting the strain field of Rayleigh-type SAWs in the weak coupling regime. Through screening the in-plane piezo-electric field with photo-generated carriers, the maximum exciton drift velocity of 600 ms−1 is reported in the scenario of type-I bandgap modulation. As a supplement, the precise steering of the exciton flux can be obtained by controlling the relative phase between the input RF excitation and exciton photo-generation.92 These results provide an important insight into the weak coupling regime between the dynamic strain fields and room-temperature excitons in 2D semiconducting systems and open the door for implementing excitonic devices into data communication and processing, sensing and energy conversion. Recently, IXs in the bilayer WSe2 exhibit a long lifetime up to ∼1 ns at room temperature, exceptionally high exciton binding energies (a few 100 meV) and gate tunability.168–170 However, the charge-neutral excitons respond weakly to the in-plane electric field and thus inhibit transport beyond the diffusion length. To overcome this drawback, SAWs were excavated to modulate the charge-neutral excitons. Fig. 5a–c show the directional transport of IXs in the bilayer WSe2 driven by the high-frequency (1.237 GHz) and 2.832 μm-wavelength SAWs. To be specified, the SAW-driven excitonic transport was activated above the threshold acoustic power at 100 K and reached 20 μm, ten times longer than the diffusion length. The temperature-dependent exciton transport measurement revealed the transition (at 30K) from the diffusion-limited regime at low temperatures to the acoustic field-driven regime at elevated temperatures. At 6 K, the IXs are highly localized and their transports are diffusion-limited. At intermediate temperatures, the IXs are bright with phonon-assisted recombination but maintain a long lifetime, enabling the SAW-driven transport over long distances. Above the Debye temperature, strong phonon scattering and decreased exciton lifetime prevent efficient transports by suppressing exciton mobility. Nevertheless, thanks to the powerful SAW modulation, a SAW-driven transport distance of ∼2 μm is still observed even at room temperature.91 The above work proves SAWs an effective and contact-free means to control the exciton dynamics and transport over a distance far beyond the diffusion length. Toward the longer transport distance, 2D semiconductors that possess natural longer exciton lifetimes and piezo-electric substrates with the outstanding acoustic velocity are preferred. This contact-free architecture preserves the material quality and prevents the formation of the local Schottky barrier and other detrimental effects from local gates.141 To date, the maximal transport distance was reported to be 20 μm at 100 K in the bilayer WSe2 and further improvements in interface cleanness and 2D material crystalline may push the SAW-involved excitonic devices operative at room temperature.
| Materials | Transport distance (L, μm) | Exciton species | Temperature (T, K) | External stimuli | Methods of generating strain |
|---|---|---|---|---|---|
| GaAs QWs131 | >265 | IXs | 12 | Dynamic strain | ZnO substrate |
| GaAs QWs132 | 1000 | IXs | 4.5 | Dynamic strain | ZnO substrate |
| GaAs QWs120 | 12 | IXs | 4 | Dynamic strain | GaAs substrate |
| WSe2 (1L)92 | 1 | DXs | RT | Dynamic strain | LiNbO3 substrate |
| WSe2 (2L)91 | 2 | IXs | RT | Dynamic strain | LiNbO3 substrate |
| WS2 (1L)133 | 0.18 | DXs | RT | Local strain | AFM tip |
| WS2 (1L)134 | 0.4 | DXs | RT | Local strain | Polymer micro-pillar |
| WS2 (1L)135 | 0.4 | DXs | RT | Uniaxial tensile | Wrinkled substrate |
| WSe2 (1L)136 | 0.5 | DXs | RT | Local strain | SiO2 pillar |
| WSe2 (1L)137 | 1 | DXs | RT | Local strain | AFM tip |
| MoSe2 (1L)134 | 1 | DXs | RT | Local strain | Polymer micro-pillar |
| MoS2/WSe2138 | 3 | IXs | RT | Local strain | Wrinkled substrate |
| MoS2 (1L)139 | 1.5 | DXs | RT | Electric field | — |
| MoS2/WSe2140 | 5 | IXs | 4.7 | Electric field | — |
| MoSe2/WSe2141 | 3.2 | IXs | 4 | Electric field | — |
| MoSe2/WSe2142 | 13 | IXs | 1.7 | Electric field | — |
| 2D perovskite (1L)143 | 0.16 | DXs | RT | None | — |
| CsPbBr3 (1L)144 | 0.2 | DXs | RT | None | — |
| MoSe2 (1L)145 | 0.4 | DXs | RT | None | — |
| WS2 (1L)130,146 | 0.36–0.75 | DX | RT | None | — |
| WSe2/MoSe2147 | ∼1 | IXs | RT | None | — |
| WSe2/hBN/MoSe2147 | 2.6 | IXs | RT | None | — |
| Lateral WSe2/MoSe2148 | 2.9 | IXs | RT | None | — |
| WSe2/PbI4149 | ∼7.6 | IXs | 78 | None | — |
| WSe2/MoSe2150 | ∼10 | IXs | 4 | None | — |
![]() | ||
| Fig. 5 Acoustic exciton transports. (a) Acoustic exciton transport in the monolayer and bilayer WSe2 on the SAW delay lines. The propagating SAWs modulate the exciton energy and transport the excitons away from the pump spot. At the flake edge, the excitons recombine and generate photoluminescence. (b and c) Real-space PL mapping (left) and spectral PL image (right) of the h-BN encapsulated bilayer WSe2 when SAWs (power ∼6 mW) are switched off (b) and on (c). The non-local exciton emission at the flake edge with an emission peak at 1.56 eV is attributed to the IXs in WSe2. After tuning on SAWs, the emission spectrum at the pump spot slightly decreases, while the emission spectrum at the flake edge drastically increases. Two bright emission spots appear at the flake edge and the focal point of two acoustic waves due to the SAW-driven transport of IXs. Reproduced from ref. 92 with permission from Springer Nature, copyright 2022. (d and e) Spatiotemporal mapping of the normalized exciton density (left) and corresponding profiles at different time instances (right) in the monolayer WSe2. The RF input powers are set to −60 dBm (d) and 13 dBm (e) at an optical fluency of 1.2 μJ cm−2. A spatial shift in the exciton density can be observed under RF excitation. At an input power of ∼−60 dBm, the exciton density symmetrically broadens in space due to exciton diffusion in the monolayer. At an input power of ∼13 dBm, an asymmetric spatial shift in the exciton density along the SAW propagation direction is observed and the observed shift increasing with time. (f) Phase-synchronized evolution of the Gaussian peak position extracted by fitting the normalized exciton density. The Gaussian peak position oscillates with the period of SAWs. A progressive delay in the Gaussian peak emerged as the time delay is increased at an increment of T/4. In addition, a net drift in the Gaussian peak aroused along the SAW propagation direction. Reproduced from ref. 91 with permission from Springer Nature, copyright 2022 (Open access). | ||
![]() | ||
| Fig. 6 Coherent acoustic phonons. (a) Setup sketch of optical SAW excitation and detection by the femtosecond time-resolved pump–probe technique. The probe beam is sent through a telescope (4f lens) system that consists of an automatically controlled two-axes rotating mirror followed by a confocal convex lens pair and a microscope objective lens, which expands the beam to fill the entire objective aperture for a diffraction-limited probe spot. The rotation axis of the scan mirror is the optical axis. (b) Sketch of a free-standing trilayer MoSe2 flake under incident laser pulses. (c) Transient reflectivity ΔR/R for 22-layer MoSe2 flakes. The filtered signal displaying the coherent acoustic B1 oscillations is shown in the top panel, and the corresponding numerical Fourier transform is depicted below, displaying a clear single peak with a frequency of ∼0.1 THz. Reproduced from ref. 171 with permission from Royal Society of Chemistry, copyright 2019. (d) Transient reflectivity ΔR/R taken at a monolayer step boundary of an 18-layer/19-layer MoS2 suspended film. (e) AFM phase-image showing the monolayer step boundary on the top of MoS2 drum (right). The horizontal dark band shows the path of 1 μm diameter beam across the surface. (f) The corresponding 2D frequency vs. the distance map showing the fast Fourier transform (FFT) spectra of the probe reflectivity, acquired at different positions while stepping the beam across the monolayer steps (e). FFT spectrum of transient reflectivity from (d) that was taken approximately at the 2.2 μm position is shown in the right panel. Reproduced from ref. 172 with permission from Springer Nature, copyright 2021 (Open access). | ||
The resonant optical excitation of infrared-active lattice vibrations is feasible to induce targeted structural deformations in 2D materials.198 In 2016, the temporal dynamics of coherent lattice vibration modes in mono- and few-layer WSe2 flakes was realized by performing time-resolved transmission measurements.199 Under ultrashort pulses resonating with A excitons of the monolayer WSe2, coherent oscillations of A1g optical phonon and longitudinal acoustic phonon at the M point of the Brillouin zone were impulsively generated. Guided by the temperature-dependent evolution of A1g optical phonons, the accelerated phonon dephasing with temperature was modeled by the anharmonic decay process, in which the A1g mode decays into two acoustic modes. In multilayer WSe2 flakes, the interlayer breathing mode (B1) is found to be sensitive to the number of layers, demonstrating its usefulness in characterizing layered TMDCs. Fig. 6b–c show the dynamics of coherent interlayer longitudinal acoustic phonons (B1 mode) for free-standing 2H-MoSe2 flakes with varying thickness from the bilayer to the bulk counterpart.171 From the temporal decay of the phonon modes (frequencies ∼1 THz for the bilayer but decrease to ∼4.2 GHz for the bulk one), it is possible to clearly distinguish an abrupt crossover related to the acoustic mean free path of phonons, and the constraints imposed on the acoustic decay channels during the reduction of film thickness. For thicker MoSe2 flakes (above ∼20 layers), the acoustic attenuations were mainly determined by the scatterings of coherent acoustic modes with thermal phonons through anharmonic three-phonon scattering processes. For flakes thinner than ∼20 layers, the predominant damping mechanism was ascribed to the surface asperity effect. In another report, the Raman-active optical phonon modes of single-layer graphene was experimentally modulated by SAW-induced dynamic strain fields.200 In particular, an effective strain field of 0.24% was generated for single-layer graphene on the LiNbO3 device at a SAW resonator frequency of ∼400 MHz. Under this strain field, the Raman scattering intensity variation of the G band as large as 15% and a phonon frequency shift of up to 10 cm−1 could be achieved. It should be highlighted that this method is universal in manipulating the optical and vibrational properties of other 2D systems.
In order to achieve intense interactions and sideband resolution (single-sideband operation), the SAWs were implemented into exciting high frequency motion and then exploring opto-mechanics with travelling acoustic waves (the regime of stimulated Brillouin scattering).201 In this way, acoustic cavities may offer spatial confinements for high frequency (1–10 GHz) elastic waves and behave as a vital part of opto-mechanical signal processing.202–206 For instance, high frequency (f) and high quality factor (Q) MoS2-based 2D acoustic cavities operating in the frequency range of 50–600 GHz with (f × Q) up to 1 × 1014 have been reported (Fig. 6d–f). In addition, well-controlled heterogeneities (monolayer steps, drum and interfaces) within the high-quality MoS2 suspended film and a frequency comb generator implemented in the h-BN/MoS2 bilayer can alter the underlying elastic-strain patterns and extend the functionality of acoustic cavities.172 These acoustic cavities in 2D heterostructures provide a pathway for exploring the quantum nature of phonons and developing phonon-based sensing and signal processing platforms.
| This journal is © The Royal Society of Chemistry 2023 |