Quantum dot infrared photodetectors: Comparison of experiment and theory

We present data and calculations and examine the factors that determine the detectivities in self-assembled InAs and InGaAs based quantum dot infrared photodetectors (QDIPs). We investigate a class of devices that combine good wavelength selectivity with “high detectivity.” We study the factors that limit the temperature performance of quantum dot detectors. For this we develop a formalism to evaluate the optical absorption and the electron transport properties. We examine the performance limiting factors and compare theory with experimental data. We find that the notion of a phonon bottleneck does not apply to large-diameter lenslike quantum dots, which have many closely spaced energy levels. The observed strong decrease of responsivity with temperature is ultimately due to a rapid thermal cascade back into the ground states. High temperature performance is improved by engineering the excited state to be near the continuum. The good low temperature $\left(77\mathrm{K}\right)$ performance in strongly bound QDIPs is shown to be due to the high gain and the low noise achievable in these micron size devices.

Figure 1

(Color online) The device structures for QDIPs (left: $D_1$ and right: $D_2$).

Figure 2

(Color online) The band diagrams for (a) $D_1$ and (b) $D_2$.

Figure 3

(Color online) Experimental photocurrent spectra for (a) $D_1$ and (b) $D_2$.

Figure 4

(Color online) Peak responsivity of (a) a device similar to $D_1$ but with a higher quantum dot doping (semilogarithm) and (b) device $D_2$ (semilogarithm).

Figure 5

(Color online) GaInAs $\mathrm{QD}∕\mathrm{Ga}\mathrm{In}\mathrm{P}∕\mathrm{Ga}\mathrm{As}$ energy levels and possible transition: (a) normal incidence, (b) $z$ polarization.

Figure 6

(Color online) InAs $\mathrm{QD}∕\mathrm{Ga}\mathrm{As}∕\mathrm{Al}\mathrm{In}\mathrm{As}∕\mathrm{In}\mathrm{P}$ energy levels and possible transition: (a) normal incidence, (b) $z$ polarization.

Figure 7

Probability of staying in the ground state with polaron correction.

Figure 8

(Color online) Noise current for (a) $D_1$ (semilogarithm) and (b) $D_2$ (semilogarithm).

Figure 9

(Color online) Square of noise current vs dark current (log-log).

Figure 10

(a) The “bias-dependent part of the responsivity” (defined as the product $N_g[{\nu }_{\mathit{ec}}{e}^{-E_{\mathit{ec}}\left(F\right)∕kT}∕({\nu }_0+{\nu }_{\mathit{ec}}{e}^{-E_{\mathit{ec}}\left(F\right)∕kT})]F=R∕R_0$) as a function of field (V/cm). (b) The “temperature-dependent part of the responsivity” (defined as $N_g[{\nu }_{\mathit{ec}}{e}^{-E_{\mathit{ec}}\left(F\right)∕kT}∕({\nu }_0+{\nu }_{\mathit{ec}}{e}^{-E_{\mathit{ec}}\left(F\right)∕kT})]F=R∕R_0$ with ${\mathrm{n}}_{\mathrm{e}}\sim 0$ and ${\nu }_0$ constant) as a function of temperature $\left(K\right)$ with polaron correction at $F=6\times {10}^4\mathrm{V}∕\mathrm{cm}$ for $N_g$ but with constant ($T$ independent) relaxation rate ${\nu }_0$.

Figure 11

(Color online) (a) Dark current of $D_1$ (semilogarithm); (b) dark current of $D_2$ (semilogarithm).

Figure 12

(Color online) Arrhenius plot of the experimental dark current of $D_1$.

Figure 13

Theoretical dark current (a) Arrhenius plot and (b) field (V/m) dependence at different temperatures.

Figure 14

(Color online) Detectivities of (a) $D_1$ (semilogarithm) and (b) $D_2$ (semilogarithm).