Why GaN and cascade devices?
There is one first major advantage to develop THz quantum cascade devices using the GaN/AlGaN material system in terms of the exceptionally broad spectral range, which can be covered using this material system. Indeed, standard III-V semiconductors like GaAs cannot be operated in the 5-11 THz frequency range because of their Reststrahlen band, i.e. the spectral region where the material is completely opaque due to the absorption by optical phonons. This is not the case for GaN because of the three fold larger energy of the optical phonons. GaN offers prospects for THz quantum cascade devices, which can operate in a much broader spectral range from 1 to 15 THz and in particular in the 5-11 THz, which cannot be covered by other III-V semiconductors (see Fig.1).
Figure 1 : Maximum operating temperature versus emission frequency of InP-
and GaAs-based quantum cascade lasers under pulsed operation. The green rectangle shows the operating
temperature and frequency range expected for nitride-based devices.
Figure 2 : Population inversion versus temperature of GaN-, ZnO- and GaAs-based
QCLs from [JAP 105, 113103, 2009]. The insets illustrate the deleterious effect of LO-phonon emission
between lasing subbands in the case of GaAs and the suppression of this effect in nitride devices.
In terms of sources, the recent demonstration of quantum cascade lasers (QCL) emitting at THz frequencies based on GaAs/AlGaAs quantum wells by a few laboratories including members of this consortium has opened new technological solutions for applications requiring a few tens-of-milliWatt power in the 1.2 to 4.6 THz spectral range. These devices rely on the ISB radiative transition in an active QW region and injector/extractor regions to transfer electrons from one period to the other. However, as shown in figure 1, the maximum operating temperature (Tmax) reported so far, 200 K (120 K) for pulsed (continuous wave) operation, has only improved marginally in the last few years and is still too low for Peltier cooling and widespread applications. One intrinsic reason limiting the Tmax is the small energy of the longitudinal optical (LO) phonon in GaAs (36 meV, 8.2 THz). Indeed, as the temperature is increased, electrons in the upper lasing subband get sufficient thermal energy for activating fast (<1 ps) non-radiative relaxations via LO-phonon emission towards the lower lasing subband, hence ruining the population inversion (see Fig. 2). It has been predicted that wide band gap semiconductor materials such as GaN, with an LO-phonon energy of 92 meV (22.3 THz), do not suffer from these parasitic non-radiative channels, paving the way for THz QCLs operating above room temperature up to a predicted Tmax of 450 K, as illustrated in Fig. 2.
In terms of QCDs, the detectivity and the maximum operating temperature can be largely improved using material with a large optical phonon energy such as GaN. Indeed, the QCD detectivity is limited by the Johnson noise governed by the device resistance at zero bias. This resistance is fundamentally limited by the parasitic transport of electrons from the ground state through the extractor or next period states mediated by the absorption of optical phonons. Since the density of optical phonons at a given temperature decreases exponentially with the phonon energy, large optical phonon energy material such as GaN are expected to provide very large R0 values and consequently enhanced operating temperatures. Estimations show that GaN QCDs operating at 5 THz frequency could operate at liquid nitrogen temperatures (77 K) to be compared to 5-10 K for the best THz detectors based on GaAs.