Red-Shifted Excitation and Two-Photon Pumping of Biointegrated GaInP/AlGaInP Quantum Well Microlasers

Abstract

Biointegrated intracellular microlasers have emerged as an attractive and versatile tool in biophotonics. Different inorganic semiconductor materials have been used for the fabrication of such biocompatible microlasers but often operate at visible wavelengths ill-suited for imaging through tissue. Here, we report on whispering gallery mode microdisk lasers made from a range of GaInP/AlGaInP multi-quantum well structures with compositions tailored to red-shifted excitation and emission. The selected semiconductor alloys show minimal toxicity and allow the fabrication of lasers with stable single-mode emission in the NIR (675-720 nm) and sub-pJ thresholds. The microlasers operate in the first therapeutic window under direct excitation by a conventional diode laser and can also be pumped in the second therapeutic window using two-photon excitation at pulse energies compatible with standard multiphoton microscopy. Stable performance is observed under cell culturing conditions for 5 days without any device encapsulation. With their bio-optimized spectral characteristics, low lasing threshold, and compatibility with two-photon pumping, AlGaInP-based microlasers are ideally suited for novel cell tagging and in vivo sensing applications.

Figure 1

(a) Wavelength-dependent absorption in blood (dashed line) and skin (solid line), data from ref (17). The shaded regions show the NIR I and NIR II “therapeutic windows”, and the colored lines indicate different pump lasers that are compared in this work. (b) Schematic of the band gap and composition of GaInP quantum wells (light gray) with AlGaInP barriers (dark gray). The colored lines illustrate the evolution of the band gap E_G for each of the four tested wafers (A, B, C, and D). (c) Photoluminescence spectra of the four wafers, color-coded according to (b). (d) Peak photoluminescence wavelength versus gallium content of the active layer for each structure.

Figure 2

Laser fabrication and characterization. (a) Microdisk lasers supported on columns of the sacrificial layer after the HBr/Br₂ etching step and prior to underetching by HF. Scale bar, 2 μm. (b) Emission spectra of a microdisk laser under 532 nm excitation. Pump pulse energy increases from purple to yellow. (c) Maximum output intensity (gray squares, left axis) versus pump pulse energy. The fit to the rate-equation model (black line) indicates a lasing threshold of 0.489 pJ. The linewidth of the lasing peaks (blue triangles, right axis) narrows near the lasing threshold and subsequently broadens at higher pump powers.

Figure 3

Bulk photoluminescence spectra for Wafers A, B, C, and D (in color, right axis) and histograms of lasing wavelength for microlasers fabricated from each wafer (left axis, n ≈ 150 for each wafer).

Figure 4

Dependence of lasing threshold on pump wavelength. (a) Normalized lasing thresholds of microlasers for different pump wavelengths. Error bars indicate the standard deviation. (b–d) Threshold characteristics at 642  nm, 10  MHz pumping for microdisk lasers made from Wafer B (b), Wafer C (c), and Wafer D (d). Thick lines in each panel represent threshold behavior predicted by the model for the average of the fitting parameters from each laser. (e) Statistics of a lasing threshold for pumping with 642 nm (pink) and 532 nm (green) light, showing mean threshold (center line), the 25–75 percentile (box), and 1.5 times the standard deviation (whiskers). Asterisk marks the sample where no lasing was detected.

Figure 5

(a) Threshold curves and (b) stability of laser emission under continuous pumping in an aqueous environment. (c) Epifluorescence microscopy image of HEK293 cells incubated with microdisk lasers for 3 days. Staining with cell tracker dye (green, live cells) and propidium iodide (magenta, dead cells). Scale bar, 40 μm. (d) Statistical analysis of cell staining showing no decrease in the viability of cells incubated with microlasers for 3 and 5 days compared to control cultures without microlasers. Mean (center line), 25–75 percentile (box) and 1.5 times the standard deviation (whiskers). (e) Normalized lasing spectra of intracellular lasers after 3 days of incubation. (f) Brightfield microscope image of HEK293 cells incubated with microdisk lasers. Arrows point to intracellular microlasers with color-coding corresponding to spectra in (e). Scale bar, 50 μm.

Figure 6

(a) Lasing spectra under two-photon excitation with power increasing from purple to yellow. (b) Input–output characteristics for several microdisks prepared from Wafer C (cyan) and Wafer D (magenta) and corresponding fits to the modified rate-equation model. (c) Rescaled threshold curves under 532 nm single-photon (squares, dashed line) and 1030 nm two-photon excitation (dots, solid line) from Wafer D microlasers. The pump rate was calculated from the linear (single-photon) and power-squared (two-photon) dependence of the excited fluorescence, respectively. The curves were normalized such that at the threshold, pump rate and output power are 1.