Carrier Decay and Diffusion Dynamics in Single-Crystalline CdTe as Seen via Microphotoluminescence

The ability to spatially resolve the degree to which extended defects impact carrier diffusion lengths and lifetimes is important for determining upper limits for defect densities in semiconductor devices. We show that a new spatially and temporally resolved photoluminescence (PL) imaging technique can be used to accurately extract carrier lifetime values in the immediate vicinity of dark-line defects in $\mathrm{CdTe}/\mathrm{MgCdTe}$ double heterostructures. A series of PL images captured during the decay process show that extended defects with a density of $1.4\times {10}^5{\mathrm{cm}}^{-2}$ deplete photogenerated charge carriers from the surrounding semiconductor material on a nanosecond time scale. The technique makes it possible to elucidate the interplay between nonradiative carrier recombination and carrier diffusion and reveals that they both combine to degrade the PL intensity over a fractional area that is much larger than the physical size of the defects. Carrier lifetimes are correctly determined from numerical simulations of the decay behavior by taking these two effects into account. Our study demonstrates that it is crucial to measure and account for the influence of local defects in the measurement of carrier lifetime and diffusion, which are key transport parameters for the design and modeling of advanced solar-cell and light-emitting devices.

Figure 1

80-K PL spectrum under 2.3-eV excitation of $45\mathrm{W}/{\mathrm{cm}}^2$. The vertical axis is expanded for the region of the low-energy band.

Figure 2

Time-integrated PL image under $700\mathrm{nJ}/{\mathrm{cm}}^2$ pulsed wide-field illumination filtered to pass both spectral bands. Bright points of low-energy PL are seen in the centers of some dark spots (arrows), indicating the system spatial resolution.

Figure 3

The squared FWHM of the Gaussian fit to the exciton PL profile as a function of time after excitation by a $3\mu \mathrm{J}/{\mathrm{cm}}^2$ focused pulsed laser. Points are measured widths; solid line is a linear fit to points. Inset: Streak camera trace from which the spatial profiles are extracted at 1-ns intervals. Solid curves show the profile through the dashed line at 7 ns and its Gaussian fit.

Figure 4

Time-resolved PL maps of the exciton PL taken 0.5 ns (a), 5 ns (b), 10 ns (c), and 25.5 ns (d) after the laser pulse. Excitation conditions and sample region are identical to Fig. 2. (e) Spatially integrated PL in a $15\text{−}\mu \mathrm{m}\text{−}\text{wide}$ bright region as a function of time for the full data set.

Figure 5

Calculated PL intensities after pulsed wide-field illumination. Times are 1 ns (a), 5 ns (b), 10 ns (c), and 25 ns (d). Model defects are a $2\text{−}\mu \mathrm{m}\text{−}\text{diameter}$ circle at ($-20$, 0) and a $2\text{−}\mu \mathrm{m}\text{−}\text{wide}$ wide rectangle at $(0,0)\mu \mathrm{m}$. Solid lines are a profile of the PL along $y=0$ drawn relative to the panel bottom.

Figure 6

Time profiles of the calculated results in Fig. 5 at points along the $y=0$ axis. Solid lines are taken at (from top to bottom) $x=-40$, $-30$, $-20\mu \mathrm{m}$, and 0. Dashed lines are exponential fits to the region 100–200 ns yielding the time constants shown. Inset: Expanded plot of the early time region comparing the 0 and $-40\mu \mathrm{m}$ profiles (lines) with normalized PL (symbols) from Fig. 4 and from a similar region of interest centered on a thin line defect.

Figure 7

(a) Time-integrated image of both PL spectral bands under focused pulsed excitation at $(0,0)\mu \mathrm{m}$, taken in a sample region that contains a bright point emitter at $(-21,0)\mu \mathrm{m}$. Inset: The same data on a rescaled gray scale map to correctly show the excitation point profile more clearly. Spatial scales are unchanged. (b) Streak camera trace of the same signal in (a).