emitters have been widely used in diverse mid-IR imaging and spectroscopic systems,
opening enormous opportunities for industrial, environmental, medical, defense and
security sensing applications3-5. Continued efforts further demonstrate that these mid-IR
systems can provide an additional dimension of contrast, such as revealing camouflaged
surfaces or the compositional and functional properties of chemical species, if the
polarization states and spectral characteristics of exploited mid-IR light can be actively
controlled.6 In addition to sensing applications, it is notable that mid-IR light sources are
at the centre of modern free-space optical communication systems, and their information
capacities can be further improved via the polarization- and spectral-encoding of optical
signals7,8. However, despite promising advances, manipulating the spectral and
polarization properties of mid-IR light generally requires the cascade of filters, dispersive
optics and polarization optics together with mechanical moving parts. These mid-IR optical
components are relatively more expensive and less developed than their visible or near-
infrared counterparts. More critically, the requirements of multiple optical and mechanical
elements pose fundamental limitations for realizing miniaturized systems with robust
integration and high-speed operation.
To circumvent these technological obstacles, one possible approach is to develop mid-IR
emitters with electrically controllable polarization and spectral properties. However,
conventional mid-IR emitters generally rely on III–V or II–VI semiconductors1,5.
Emissions from these three-dimensional materials are typically nonpolarized. Wavelength-
tuneable emitters with electric control can in principle be achieved by vertically stacking
two or more electroluminescence (EL) units (i.e., creating tandem structures). These
emitters, operated within the visible region, have been successfully demonstrated using
solution-processed organic polymers9,10. Unfortunately, these wavelength-tuneable
devices require multilayered architectures to engineer the band profiles of heterostructures
and multiple electrodes to independently control each EL unit. Exploiting epitaxial
semiconductors, which generally suffer from lattice and thermal mismatches at
heterointerfaces1,5, to form tandem LEDs is a formidable challenge.
In these respects, exploiting a family of layered van der Waals (vdW) materials might be a
promising alternative route for developing electrically controllable mid-IR sources. One
prominent virtue offered by this material family is that different vdW materials can be
stacked vertically with arbitrary chosen sequences and crystal orientations due to their
weak vdW interactions11-13. Moreover, vdW materials provide a large variety of optical
bandgaps, spanning across the electromagnetic spectrum from ultraviolet to terahertz14,15.
For mid-IR emitter applications, black phosphorus (BP) has gained enormous attention,
because it has a direct and narrow gap (~0.33 eV), low Auger recombination characteristics
and an in-plane anisotropic structure16-22. By leveraging these exotic features, pioneering
works have not only successfully demonstrated BP-based mid-IR LEDs23-25, exhibiting
characteristics of linear polarized emission and high emission efficiency, but have also
shown their great promise for integrated mid-IR silicon photonics and gas sensing
applications23,26. In addition, doping BP with arsenic to form black arsenic phosphorus (b-
AsP) results in shrinkage in the band gap16,20,27. Applications of b-AsP to mid-IR