Terahertz Sources Based on Emission from a $\mathrm{Ga}\mathrm{As}$/($\mathrm{Al}$,$\mathrm{Ga}$)$\mathrm{As}$ Heterostructure at Cryogenic Temperatures

A high-electron-mobility ${\mathrm{Ga}\mathrm{As}/\mathrm{Al}}_{0.36}{\mathrm{Ga}}_{0.64}\mathrm{As}$ heterostructure is processed by electron-beam lithography to fabricate samples with a separation of the current-supplying Ohmic contacts of between $28\mu \mathrm{m}$ and 1.3 mm. Voltage pulses of 20-ms duration, with a 4% filling factor and an amplitude up to 30 V, stimulate terahertz emission from samples cooled to 4 K. The radiation spectrum, centered at about 390 GHz (about 1.6 meV), registered with a Fabry-Perot interferometer, shows almost no dependence on the applied voltage or on the shape of the sample. The emission appears in a threshold manner and is accompanied by a jump of current by more than 2 orders of magnitude. The energy of the emitted photons is about three times smaller than in the case of any previously reported impurity-related emission from either bulk $\mathrm{Ga}\mathrm{As}$ or $\mathrm{Ga}\mathrm{As}$-based quantum wells. The emission is interpreted as resulting from a two-step ionization-recombination process involving shallow donors in the ($\mathrm{Al}$,$\mathrm{Ga}$)$\mathrm{As}$ barrier, which create bound states of electrons in the $\mathrm{Ga}\mathrm{As}$ quantum well. Thus, the frequency of emission is argued to be tuned by the width of the spacer layer. We present these as small convenient sources that can be used for calibration and check-up operation of cooled bolometric systems used in astronomical observations.

Figure 1

A schematic of the experimental setup. The inset (not to scale) shows a top view of the samples studied, with dimensions of the current-conducting channels as indicated.

Figure 2

(a) The spectra of the radiation emitted from sample ${\mathrm{S}}_1$ at different amplitudes of the voltage pulses: 20 V (bottom) and 25 V (top). (b) The spectra of the radiation emitted from sample ${\mathrm{S}}_2$ at different amplitudes of the voltage pulses: 10 V (bottom) and 15 V (top). The spikes marked 215 GHz and 630 GHz are the spectra of monochromatic radiation.

Figure 3

The $I$-$V$ characteristics of sample ${\mathrm{S}}_2$ (left scale) and the intensity of the terahertz emission (right scale). The gray rectangle describes the noise level of the detector. The inset shows the dependence of the threshold voltage on the magnetic field for sample ${\mathrm{S}}_2$.

Figure 4

The spectra of the radiation emitted from sample ${\mathrm{S}}_2$, biased with voltage pulses, with the amplitude equal to 12 V measured at different magnetic fields.

Figure 5

(a) A schematic of the top layers of the heterostructure studied, including a $\mathrm{Ga}\mathrm{As}$ channel, a ($\mathrm{Al}$,$\mathrm{Ga}$)$\mathrm{As}$ barrier with two $\delta$-doping layers, and a $\mathrm{Ga}\mathrm{As}$ cap layer. The separation of the $\delta$ layers is 30 mm. (b)–(d) The results of the self-consistent Schrödinger-Poisson calculations, showing the energy of the bottom of the conduction band in the region of the triangular quantum well for $N_A$ equal to ${10}^{14}{\mathrm{cm}}^{-3}$, $2\times {10}^{15}{\mathrm{cm}}^{-3}$, and $8\times {10}^{15}{\mathrm{cm}}^{-3}$ (b)–(d), respectively). The red horizontal bars show the energy of two lowest electron levels in the quantum well. The horizontal scale shows the distance to the top surface of the heterostructure.

Figure 6

The dependence of the binding energy of a donor on its position in a ($\mathrm{Al}$,$\mathrm{Ga}$)$\mathrm{As}$/$\mathrm{Ga}\mathrm{As}$/($\mathrm{Al}$,$\mathrm{Ga}$)$\mathrm{As}$ quantum well in the case of a finite potential barrier (0.318 eV) and a well width equal to 20 nm. The origin of the horizontal axis is in the middle of the quantum well. The thin vertical line shows the position of the $\mathrm{Ga}\mathrm{As}$/($\mathrm{Al}$,$\mathrm{Ga}$)$\mathrm{As}$ interface. The red points from 0 nm to 30 nm are from Ref. [22]; a red point at 58 nm corresponds to the analysis presented in the present work. The error bar is equal to half of the full width at half maximum of the emission peak for sample ${\mathrm{S}}_1$.