Introduction

Cell labeling techniques play an essential role in cell imaging and individual cell tracking1. In medical and biological science, cell labeling is used to study cell migration2 and differentiation3, the dynamic behavior of cells4, and the progression of diseases, including the metastasis of cancer5. Fluorescent dyes have been widely used for cell labeling, but their broad emission spectrum has been an obstacle to labeling large numbers of cells6,7,8. In addition, issues such as the low efficiency and photobleaching of organic materials remain to be resolved8. A new approach using intracellular lasers has been proposed to address these issues9,10,11,12. The few hundred-fold narrower linewidth and higher signal-to-noise ratio of intracellular laser probes demonstrated the advantages of using intracellular lasers over fluorescent dyes for cell labeling and tracking applications. In particular, inorganic optical materials13,14,15,16,17,18,19 have been evaluated for intracellular laser application because they offer excellent lasing characteristics and chemical and mechanical stability20,21,22,23,24. For example, inorganic materials, including cadmium sulfide (CdS) nanowires22, aluminum gallium phosphide (AlGaP) multiquantum well nanodisks21, and indium gallium arsenide phosphide (InGaAsP) nanodisks20, have been used for inorganic intracellular lasers to exploit their advantages. While the toxicity of CdS and InGaAsP is an obstacle to the bioapplication of these compounds and the toxicity of aluminum gallium indium phosphide (AlGaInP) is not well known25, gallium nitride (GaN) is a well-known, nontoxic, biocompatible material26,27 with high refractive index and optical gain coefficient28. Here, we report the fabrication of GaN microrod lasers and their lasing characteristics for intracellular laser applications.

Materials and methods

GaN microrod growth by CVD on graphene

GaN microrods were grown on graphene films by metal-organic vapor phase epitaxy (MOVPE), as shown in Fig. 1a. Graphene films were synthesized on copper foil using chemical vapor deposition (CVD) and transferred onto supporting substrates of amorphous SiO2-coated Si (SiO2/Si). Typically, CVD-grown graphene films are semitransparent, multilayer graphene with electrical resistance in the range of 600–800 Ω/sq. GaN microrods were grown on graphene films using a two-step temperature process: 750–850 °C for 3 min and 950–1050 °C for 30 min. The substrate was then heated at 1100 °C for 10 min with hydrogen. Trimethyl gallium (TMGa), ditertiary butyl silane (DTBSi), and ammonia (NH3) were employed as reactants to grow the GaN microrods, and nitrogen was used as the carrier gas. The flow rates of TMGa, DTBSi, and NH3 were in the ranges of 15–30, 1–3, and 100–500 sccm, respectively. The pressure of the reactor chamber was maintained at 300 Torr with a mixture of hydrogen and nitrogen gases during the process. The GaN microrods grown over the entire graphene film had a uniform areal density of 107 cm−2.

Cell culture

HeLa cells were cultured in modified Eagle’s medium supplemented with 10% fetal bovine serum, 1% GlutaMAX, and 1% penicillin/streptomycin and maintained at 37 °C in a humidified atmosphere with 5% CO2.

Internalization of GaN microrods into cells

The as-grown GaN microrods on the graphene/SiO2/Si substrate were sterilized in 95% ethyl alcohol for 1 h, and then, GaN microrods were detached from the substrate in 1 ml of cell culture medium by sonication for 1 min. The concentration of GaN microrods in the cell culture medium was adjusted to 107 ml−1, and then, 200 μl was added to a cell culture dish using a micropipette. GaN microrods were naturally internalized into the cells via endocytosis a few hours after being added to the cell culture medium.

Fluorescence imaging

Calcein AM and 4′-6-diamidino-2-phenylindole (DAPI) were used to label the cytoplasm and nucleus, respectively. Fluorescence images were obtained using an inverted microscope (Olympus IX73) equipped with an electron-multiplying charge-coupled device (EMCCD) camera (Andor) and an XYZ automated stage (ASI). For epifluorescence excitation, cells were illuminated with a white LED (Lumencor) and visualized with a ×40/0.6 NA objective (Olympus). Fluorescence signals were collected by the same objective and delivered to an EMCCD camera after passing through a dichroic mirror and appropriate filter sets (Chroma 49002 ET-EGFP filter set for calcein AM and 49000 ET-DAPI filter set for DAPI).

Optical setup for lasing experiments

Lasing experiments were conducted with microphotoluminescence (µ-PL) spectroscopy at room temperature. The third harmonic (355 nm) of a Nd:YAG laser (10 Hz, 10 ns pulse width) was used as the excitation source, and the laser beam was focused to a spot size of 8 μm through a UV objective lens (NA = 0.50). A conventional motorized stage was used to move the sample, and emitted lasing signals were analyzed with a conventional monochromator and a charge-coupled device (CCD) camera.

Results and discussion

Schematic illustrations of the experimental procedure are shown in Fig. 1a. We used the experimental method described in our previously reported conference paper29. GaN microrods were prepared on CVD-grown graphene films using MOVPE without metal catalysts30. As shown in Fig. 1b, the fabricated GaN microrods typically had a diameter of 1.5 ± 1 µm and a length of 10 ± 3 µm. Microrods that were detached from the substrate using sonication were added to HeLa cell cultures, as shown in Fig. 1c. Optical microscopy (OM) revealed bright spots (indicated by arrows) corresponding to the GaN microrods. HeLa cells internalized the microrods by endocytosis within a few hours after the microrods were added to the culture medium. Intracellular lasing spectra were obtained by confocal µ-PL spectroscopy.

Fig. 1: Schematic diagram of the procedure for intracellular laser experiments.
figure 1

a Schematic illustrations of the procedure for intracellular GaN lasing measurement. b Field emission scanning electron microscopy of GaN microrods grown on a graphene film. c An OM image of HeLa cells with intracellular GaN microrods.

Full size image

We investigated the endocytosis of a microrod into a cell before the lasing experiments. Representative time-lapse images (Fig. 2a) show that a HeLa cell was initially migrating near a GaN microrod for a few hours. After 5 h, filopodia and lamellipodia extended to the microrod, as indicated by the arrow in Fig. 2a. The GaN microrod was fully internalized into the cell within several minutes. We also evaluated the internalization percentage using OM images. We found that more than 70% of cells contained GaN microrods out of 53 cells (Fig. S1). Because the GaN microrods have a high aspect ratio, they can be more easily internalized into cells than the particles that have the same volume but a low aspect ratio31,32. To confirm that GaN microrods are internalized, three-dimensional (3D) fluorescence images were obtained by confocal laser scanning microscopy (CLSM). As shown in Fig. 2b, the Z-projection and two cross-sectional images along the axial and radial directions show that GaN microrods (outlined with white dashed lines) are observed in the cytoplasm. These results indicate that the GaN microrods were fully internalized into the cytoplasm.

Fig. 2: Endocytosis of GaN microrod into HeLa cells.
figure 2

a Time-lapse images of the endocytosis of a GaN microrod into a HeLa cell. b Z-projection (top) of and cross-sectional (center and bottom) fluorescent CLSM images with maximum intensity. Blue and green indicate the nucleus and cytoplasm, respectively.

Full size image

The cellular activity was monitored using phase-contrast optical microscopy (OM) to evaluate the biocompatibility of the GaN microrods. Figure 3a shows representative time-lapse images of cell migration and normal cell division over 2 weeks in culture. The HeLa cell that contained two GaN microrods in Fig. 3a migrated randomly. After 5 h, the cell entered mitosis and appeared to round up. The daughter cells are positioned in the adhesion pattern of the mother cell. This cell division appearance is consistent with normal cell division33. Following cell division, the internalized microrods (indicated by white arrows) remained in one of the daughter cells.

Fig. 3: Cell viability checking.
figure 3

a Time-lapse images of a HeLa cell undergoing normal cell division after internalizing a GaN microrod. The cell indicated by the red dashed line divided into two cells indicated by the orange and green lines. b Phase-contrast and fluorescent optical microscopy images of HeLa cells. Fluorescent images were obtained using calcein AM and ethidium homodimer dyes.

Full size image

The cell viability was evaluated using a commercial assay. As shown in Figs. 3b and S2, fluorescence imaging of calcein acetoxymethyl (AM) was observed in all cells, indicating that they were viable. In contrast, no fluorescence was observed with ethidium homodimer staining in the same culture dish, indicating that no cells were dead. Superimposing the images from phase-contrast and calcein AM and ethidium homodimer fluorescence imaging confirmed that each cell imaged was viable.

GaN microrods in cells were investigated at room temperature using confocal µ-PL spectroscopy. Figure 4a shows the μ-PL spectra of the GaN microrods at excitation power intensities ranging from 20 to 720 kW/cm2. The inset shows an OM image of the GaN microrods used for these measurements. Below the lasing threshold (Ith = 270 kW/cm2), a broad emission centered at 370 nm was observed around the near-band-edge (NBE) emission of GaN. As the excitation power intensity increased, additional sharp peaks appeared in the NBE emission spectra and eventually became the dominant feature in the PL spectra. From the mode spacing, these sharp peaks were attributed to Fabry-Perot (FP)-type lasing along the axial direction of the microrods. According to previous reports of GaN microrod lasing on SiO2 substrates, the mode spacing for FP resonances is given by \({\Delta}\lambda \sim \frac{a}{L}\), where Δλ is the mode spacing, L is the cavity length, and \(a = 0.014\) µm2 is the value experimentally obtained by linear fitting between Δλ and 1/L23. For the lasing spectra shown in Fig. 4a, Δλ is approximately 1.11 nm, which is in good agreement with the value calculated from the above equation when L = 12.3 μm, which is close to the length of the microrods. We also calculated the length dispersion of the GaN microrods from the mode spacing in Fig. S3. The average length of the GaN microrods is 11.5 µm with a standard deviation of 2.4 µm. This calculated length dispersion of the GaN microrods from Fig. S3 is in good agreement with the length of the GaN microrods measured from the SEM image in Fig. 1b. Additionally, the measured Δλ is substantially distinct from the WGM mode spacing estimated for the microrods, which is ~10 nm for a microrod with a diameter of 1.5 µm, excluding the possibility of WGM lasing. Furthermore, the average Q-factor for the observed laser emission was estimated as 620 from the observed average linewidth of 0.6 nm, which is a reasonable value for laser emissions from GaN microrods under intercellular conditions. The lasing spectra depend on the geometry of the microrods. Because the lasing exhibits Fabry-Perot type oscillation, the mode spacing depends on the length of the microrods. Additionally, the spectra could be influenced by the diameter, cross-sectional shape, or end-facet morphology. The diameters of the GaN microrods are larger than the single-mode diameter; therefore, transverse waveguide modes could be excited simultaneously in a single microrod. The existence of multiple transverse modes and competition between them resulted in complicated lasing spectra with irregular mode spacings and intensity ratios, as shown in Fig. 4a depicting a PL spectrum for an excitation power of 3.98 Ith.

Fig. 4: Characteristics of intracellular GaN microrod laser.
figure 4

a PL spectra measured at different optical pumping power densities. The inset shows an OM image of the intracellular microrod used for the measurement. b Plot of output PL intensity versus input excitation power.

Full size image

Figure 4b shows a plot of the integrated PL intensity versus excitation pump intensity. At excitation power levels below the lasing threshold, the slope is 3.7, but above the lasing threshold, the slope increases to 10.4. The threshold value of 270 kW/cm2 is similar to those reported for high-quality GaN micro- and nanostructures grown on single-crystal substrates34,35,36, demonstrating that the GaN microrods grown on graphene films are of high optical quality. Additionally, due to the relatively high refractive index of GaN microrods (nGaN = 2.6)37, the lasing threshold is low under intracellular conditions (ncell = 1.36)38. The threshold value is also comparable to that in previously reported papers on inorganic intracellular lasers20,21. Because the GaN microrods have a comparable lasing threshold value along with nontoxicity and biocompatibility, GaN microrods can be candidates for intracellular laser applications.

We investigated the individual lasing characteristics of intracellular GaN microrod lasers. Figure 5a, b show OM images of cells with internalized microrods and the corresponding lasing spectra, respectively. The lasing spectra from individual microrods are distinguishable by the center peak wavelength, mode spacings, and relative PL intensities between modes. In other words, each microrod has a distinct lasing spectrum, so each microrod can be tracked individually, making the GaN microrods suitable for cell labeling. Moreover, the effect of a refractive index change in the cell environment was similar for different transverse modes, suggesting that the profile of the laser emission spectra will not change substantially due to refractive index changes, as shown in Fig. S4. This property can be advantageous for cell tagging because the lasing spectra of individual microrods distinguished by multiple lasing modes can still be sustained under the condition of a refractive index change of the cell.

Fig. 5: Lasing spectra from different intracellular microrod lasers in living cells.
figure 5

a OM images of cells containing GaN microrods. Microrods are indicated by white arrows. b Lasing spectra obtained from the intracellular microrods shown in panel (a).

Full size image

After conducting the lasing measurements, the cell viability was tested (Fig. S5). Even when HeLa cells were exposed to a focused laser beam at 10 mJ/cm2 for 2 s, the cells remained alive. Furthermore, cell nuclei were less exposed to the focused laser beam because the GaN microrods were typically positioned at a distance from the cell nucleus39,40.

Conclusion

In summary, we fabricated GaN microrod lasers for intracellular labeling applications. The GaN microrods grown on graphene films using MOVPE exhibited excellent lasing characteristics. The use of graphene as a substrate enabled facile detachment of the GaN microrods from the substrate and dispersion into culture media for internalization into cells. The high aspect ratio of the GaN microrods allowed them to be internalized into cells more easily. Excellent lasing signals were observed under intracellular conditions due to the optical quality and high refractive index of the GaN microrods. However, UV light can be inappropriate for long-term experiments and in vivo imaging. To avoid the use of a UV light source, our future plan is to fabricate an InGaN nanorod intracellular laser. By changing the In composition, we can tune the light emission up to the visible range. We believe that intracellular GaN-based microrod lasers can be candidates for cell labeling and tracking applications.