Site-selective enhancement of Eu emission in delta-doped GaN

Abstract

Europium-doped gallium nitride (GaN:Eu) is a promising platform for classical and quantum optoelectronic applications. When grown using organometallic vapor-phase epitaxy, the dominant red emission from Eu exhibits an inhomogeneous photoluminescence (PL) spectrum due to contributions from several non-equivalent incorporation sites that can be distinguished with combined excitation emission spectroscopy. Energy transfer from the GaN bandgap to the majority site is inefficient, limiting the performance of GaN:Eu LEDs and resulting in an inhomogeneous emission spectrum dominated by disproportionate contributions from minority sites. In this work, we use site-selective spectroscopy to characterize the photoluminescence properties of delta-doped structures with alternating doped and undoped layers of varying thicknesses and demonstrate that they selectively enhance emission from the majority site when compared to uniformly-doped samples. Samples with 2-nm and 10-nm doped layers show much greater PL intensity per Eu concentration as well as more efficient energy transfer to the majority site, which are both highly desirable for creating power-efficient LEDs. Meanwhile, a sample with 1-nm doped layers shows emission only from the majority site, resulting in a narrow, homogeneous emission spectrum that is desirable for quantum technologies. This utilization of delta-doping has the potential to be broadly applicable for engineering desirable defect properties in rare-earth doped semiconductors.

Figures (5)

• Figure 1: Europium incorporation sites in uniformly-doped GaN. (a) Combined excitation-emission spectroscopy (CEES) of a uniformly doped GaN:Eu sample showing distinct incorporation sites labeled with dashed lines to indicate relevant regions. Two distinct peaks in excitation are shown for OMVPE4 due to coupling to different phonons. (b) Emission spectra of the three dominant sites obtained by spectral decomposition of the CEES map. Extracted spectra are shown at their peak amplitudes at energies noted by colored arrows in (a). (c)Composite emission spectrum obtained by summing the spectra in (b) (bottom, orange) compared to measured emission spectrum of the same sample when excited above bandgap with a 351 nm laser (top, blue).
• Figure 2: Site-selective spectroscopy of uniformly doped (UD) and delta-doped (DD) GaN:Eu. (a)-(d) Cross-sectional schematics of GaN:Eu samples studied in this work. All samples are grown on a c-plane sapphire substrate with a 30-nm-thick low-temperature (LT) buffer layer and a 2-µm-thick undoped GaN layer before the Eu-doped region. (a) is the same uniformly-doped sample depicted in Fig. \ref{['fig:sites']} with a 300nm thick doped region, while (b)-(d) contain 40 pairs of alternating doped and undoped layers of different thicknesses. (e)-(h) (Above) Resonant CEES maps, plotted using a shared colormap with thresholded counts such that the minority sites are clearly visible. (Below) Composite emission spectra calculated by summing the OMVPE1/2, OMVPE4, and OMVPE7 components obtained via spectral decomposition. Amplitudes are normalized to the peak of the brightest sample. (i)-(l) Emission spectra of each sample taken with near-UV excitation at 351 nm (above bandgap, lower row) and 364 nm (below bandgap, upper row).
• Figure 3: Eu concentration and resonant brightness of GaN:Eu samples. (a) SIMS counts per frame detected for $\mathrm{~^{151}Eu}$ over a 50 $\mu$m field of view as a function of image frame, rebinned every 10 frames to reduce noise. Each frame corresponds to an etched depth of 0.5-0.6 nm. (b) Integrated PL brightness ($I_{\mathrm{PL}}$) of OMVPE2, OMVPE4, and OMVPE7 extracted from CEES maps using spectral decomposition. Values are rescaled according to the measured absolute brightness of OMVPE4 of each sample, with error bars corresponding to the standard deviation of counts over a spatially-resolved photoluminescence scan. (c) Total number of SIMS counts ($\sigma_{\mathrm{SIMS}}$) for each sample, taken by summing over the shaded region in (a). Error bars correspond to the standard deviation of rebinned data points. (d) Concentration-normalized photoluminescence intensity, $I_{\mathrm{PL}}/\sigma_{\mathrm{SIMS}}$.
• Figure 4: Temperature dependence of the emission spectra of the four samples taken with above-bandgap (351 nm) excitation, with dashed lines indicating the effective zeros for each curve.
• Figure 5: Temperature dependence of the total integrated intensity of the $~^5\!D_0 \rightarrow ~^7\!F_2$ emission. For each sample, the data are normalized to the intensity at low temperature. (a) compares the total intensity for all four samples under above-bandgap (351 nm) excitation. (b) compares the total intensity of the 10:1 DD sample under above bandgap excitation (red) and under phonon-assisted $~^7\!F_0 \rightarrow ~^5\!D_0$ excitation (blue).