Introduction

Holography provides an approach to record and reconstruct three-dimensional (3D) wavefronts of objects1, which can be implemented in various fields, including displays, imaging2, data storage3, optical communications4, information encryption5, and neuromorphic photonic computing6. To facilitate the practicality and adaptability of these applications in chip-scale, recent advancements have focused on miniaturization and dynamic switching capability of holographic systems7,8. Spatial light modulators (SLMs) and digital micromirror devices (DMDs), combined with external laser sources, have been traditionally employed for dynamic wavefront manipulation. Nonetheless, both the complex addressing circuits embedded in these digital holographic devices and the bulky size of lasers impose fundamental challenges to achieve on-chip integration. This hinders the development of emerging wearable and portable holographic information processing products.

To address this problem, incorporating integrated light sources with diffractive optical elements (DOEs) has been proven as a promising compact and versatile photonic systems9. In this context, vertical-cavity surface-emitting lasers (VCSELs) have attracted tremendous attention10,11,12. Such semiconductor lasers, with an epitaxial active layer and distributed Bragg reflectors (DBRs) to form the laser gain medium and resonant cavity, feature unique advantages including micron-scale size, two-dimensional (2D) array format, planar emitting surface, and high modulation speeds (above GHz). By further leveraging advanced micro- and nano-fabrication techniques (such as electron beam lithography and 3D laser nano-printing), optically digitalized holograms13 can be directly integrated onto the emitting surface of VCSELs to modulate the wavefront and the polarization states of output beams14,15,16,17, which can significantly reduce the size of holographic systems. Unfortunately, as the overall area of hologram is limited by the small emitting surface of VCSELs, the sub-micro- or nano-scale pixel resolution hinder the layout of addressing circuits to obtain an optoelectronic device for dynamic light-field control.

Other pioneering methods to dynamically manipulate the optically digitalized holograms have also been proposed, such as chemical reactions18, optical erasure and rewriting19, mechanical stretching20, twisting cascaded holograms21, and thermal regulation22. Nevertheless, additional components or steps often slow down the switching frame rate and increase system complexity, which makes them inapplicable for VCSEL-integrated systems. Recently, the physical dimensions of light, such as wavelength23, polarization24, space25, time26, and modes27, have been utilized to increase the information channels of optically multiplexed holograms. To be specific, optically addressable dynamic holography can be realized by selectively controlling the properties of the input and output beams28. Therein, spatial modes multiplexing is particularly promising in boosting information channel capacity due to its theoretically infinite states. Notably, since the transverse dimension of the active region is much larger than its longitudinal length, multi-mode VCSELs usually operate in multiple transverse spatial modes, which can even be switched by adjusting the injection current29. As such, we exploit this fact to encode various holographic information channels into distinct transverse spatial modes from a multi-mode VCSEL and multiplex them into an optically digitalized hologram. Consequently, dynamic holography can be achieved by elaborately manipulating the injection current-dependent spatial modes.

Here, we develop a dynamic holographic VCSEL chip by integrating a multi-mode VCSEL with a current-addressed mode-multiplexed optically digitalized hologram. The principle to design such an ultra-compact holographic system with ultra-high switching frame rate is illustrated in Fig. 1. The lack of precise physical models of the output electrical fields from multi-mode VCSELs render it difficult to design the spatial mode-multiplexed hologram. To overcome this challenge, the orthogonal orbital angular momentum (OAM) states have been utilized to encode holographic images30 owning to the well-defined injection current-dependent OAM modes weighting coefficient spectrum (Fig. 1a). To minimize the channel crosstalk in optical multiplexed holography, the dominant OAM mode components characterized by the largest weighting coefficients (l1, l2, l3 and l4 in Fig. 1a) are selected in the design of the OAM-multiplexing hologram. In this regard, only a spatial mode state with a dominant OAM component (e.g., l2 in Fig. 1b), given by the multi-mode VCSEL at a specific injection current, selectively reconstructs the target holographic image with a stronger Gaussian intensity profile at each pixel. Notably, the hologram can be directly integrated onto the VCSELs’ surface using 3D laser nanoprinting (Fig. 1c), which further enables a holographic VCSEL chip comprising hologram arrays (Fig. 1d). As such, a dynamic holographic display can be achieved by varying the injected currents of each device over time sequences. Compared with other schemes5,25,28,31, our strategy illustrates a compact, all-in-one holographic system (including laser source and holographic devices) with only hundreds of microns. Moreover, the high-speed modulation capabilities of VCSELs allow theoretical frame rates exceeding 109 frames per second (Fig. 1e).

Fig. 1: Schematic diagrams of structure, design principle and performances of the holographic VCSEL chip.
Fig. 1: Schematic diagrams of structure, design principle and performances of the holographic VCSEL chip.
Full size image

a The injection current-dependent OAM modes weighting coefficient spectra of the multimode VCSEL. Four dominant OAM mode components at distinct currents (right panel) have been selected to design the OAM-multiplexing hologram. b Schematic of the selective holographic reconstruction of the OAM-multiplexing hologram using the multi-mode VCSEL at a specific injection current (7 mA). The big green circle in the left square bracket and the k-space analysis with solid lines (upper panel) reveal the largest OAM weighting coefficient at l2. c Illustration of incorporating an optically digital hologram onto the VCSEL’s surface through 3D laser nanoprinting (left panel) and the scanning electron microscope (SEM) image of the holographic system (right panel). d The conceptual illustration of ultra-compact holographic VCSEL chip for 3D dynamic holographic display. e The comparison of our work with the state of the art. Here, the holographic system includes the laser source and the holographic device, which is related to the hologram.

Results

OAM mode dynamics in multimode VCSELs

The physical mechanism of injection current-dependent OAM modes weighting coefficient spectra dynamics in multimode VCSELs is revealed in Fig. 2. Here, we explore the OAM mode components of VCSEL beams using the fiber model. Throughout this manuscript, oxidation-confined, top-emitting VCSELs emitting at ~ 905 nm have been employed, whose epitaxial structure mainly includes p-DBR, active region, n-DBR, and GaAs substrate (Fig. 2a). The fundamental properties of these VCSELs, including their spectral characteristics, I–V behavior, and lifetime, are presented in Supplementary Fig. 1–3 (“Methods”), respectively. In this context, the oxide aperture diameter (d) denotes the unoxidized region in the oxidized layer. The refractive index (n₂) of the oxidized region is lower than that of the unoxidized region (n₁), forming a cylindrical waveguide structure. In this model, the unoxidized central current aperture serves as the waveguide core, while the oxide confinement layer acts as the cladding. By computing the refractive indices across all epitaxial layers, an average cladding–core index difference of approximately 0.01 can be obtained32,33. As a result, using the weak-guidance approximation, a specific OAM mode state in VCSELs can be investigated by combining degenerate odd and even LPlm modes with phase differences of π/2 or 3π/2 (i.e., OAM±l = LPlma ± i × LPlmb)34. Here, l and m denote the azimuthal (OAM order) and radial indices, while a and b indicate the degenerate odd and even LPlm modes, respectively. Notably, only the LPlm modes with m = 1 are considered owning to the degradation of higher-order LP modes in propagation. Next, represented by the LP modes with the same OAM mode order l, the OAM modes capacity of multi-mode VCSELs with different oxide aperture sizes have been analyzed in detail. For this purpose, the normalized frequency (V) determined by the oxide aperture diameter (d) in VCSELs has been calculated, which indicates the number of guided wave modes supported in such fiber model (Supplementary Note 1). Moreover, to figure out each particular guided LP modes, the longitudinal normalized propagation constant (β) representing the axial component of the light wave vector are constrained to be positive. As such, the VCSELs with larger oxide aperture sizes can support more OAM modes (Fig. 2b), laying the physical foundation for providing potential information channels in OAM multiplexing.

Fig. 2: Injection current-dependent OAM modes dynamics in multimode VCSELs.
Fig. 2: Injection current-dependent OAM modes dynamics in multimode VCSELs.
Full size image

a The detailed illustration of the VCSEL structure. b The OAM modes capacity of the multi-mode VCSELs with distinct oxide aperture diameters. c The emergency of spatial hole burning (SHB) effect at the active layer center of VCSEL with the increase of the current and the oxide aperture size. d The theoretical and experimental observation of a dominant OAM mode component switching from l = 1 to l = 2 with the increase of injection current.

To further explore the injection current-dependent OAM modes competitions in multi-mode VCSELs, current density spatial distributions in the active layer and the resultant OAM modes components at various injection currents have been both studied based on the epitaxial structure and doping concentration (Supplementary Table 1). As shown in Fig. 2c, while the charge carrier distribution exhibits a Gaussian profile at a low injection current for the VCSEL with a 4 μm-diameter oxide aperture, depleting carriers can be observed at the center with the increase of the current and the aperture size, which is termed the spatial hole burning (SHB) effect35,36. The right inset of Fig. 2c illustrates a significant SHB phenomenon, which can be attributed to the preferential current flow along the annular edges of the VCSEL with a lower electrical resistance in accordance with Kirchhoff’s law. Notably, the gains of each OAM mode are determined by continuously changing current densities, giving rise to the injection current-dependent OAM modes weighting coefficient spectra (Supplementary Note 2 and Supplementary Table 2). Simulating VCSELs with large oxide apertures is challenging due to increased computational complexity and potential modeling errors. In addition, devices with different aperture sizes are equivalent in revealing the principles of mode dynamics. Thus, without losing generality, we experimentally and theoretically investigated the weighting coefficients of OAM components with l = 1 and 2 for the VCSEL with a 6 μm-diameter oxide aperture (Fig. 2d). The theoretical results (the inset is the enlarged dashed box) suggests that the dominant OAM component switches from l = 1 to l = 2 at an injection current of ~ 10 mA, which is consistent with the experimental result.

Current-addressed mode-multiplexed holography

Implementing such current-dependent dominant OAM components for the design of an OAM-multiplexing hologram can enable current-addressed dynamic holographic displays. The schematic to measure the OAM mode components of the output electrical field provided by the VCSEL with a 30 μm-diameter oxide aperture is illustrated in Fig. 3a, wherein different spiral phase plates (SPPs) were adopted to achieve inverse OAM mode conversion (“Methods”, Supplementary Fig. 4). As an example, the intensity distributions at the focal plane are given when the injection current is 28 mA, revealing the OAM modes weighting coefficient spectrum through the integrations of preset Gaussian profile areas (Fig. 3b). Notably, the OAM mode with l = 25 is a dominant component with the largest weighting coefficient. Next, the injection current was changed within the VCSEL’s working range (Supplementary Fig. 5), leading to diverse weighting coefficient distributions of the eight OAM modes with l = 22-29 (Fig. 3c). Supplementary Movie 1 directly visualizes the evolution of the light field under stepped current modulation, revealing the injection current–dependent spatial mode behavior of the VCSEL. To reduce the crosstalk in optical multiplexed holography, we elaborately propose an enumeration algorithm to identify distinct encoding OAM modes weighting coefficient spectra that feature a distinguishing largest weighting coefficient (Fig. 3d). The algorithm involves two main steps. First, the measured OAM weighting coefficients spectra are grouped, with each group containing four OAM mode components featuring distinct dominant OAM components. Second, the groups are evaluated to determine the one where the dominant OAM components show the greatest disparity in weighting coefficients compared to the other components within each spectrum. Hence, for the spatial modes obtained at four specific working currents (I = 101, 67, 28, and 58 mA) (Supplementary Fig. 6), the OAM modes (l = 23, 24, 25, and 26) have been regarded as the dominant OAM components in our scheme (Fig. 3c).

Fig. 3: Principle of current-addressed mode-multiplexed holographic encoding.
Fig. 3: Principle of current-addressed mode-multiplexed holographic encoding.
Full size image

a Conceptual illustration of the optical setup. SPP: spiral phase plate. b The intensity distributions at the focal plane when VCSEL beams at an injection current of 28 mA illuminated on SPP with distinct OAM orders (left panel), the resultant OAM modes weighting coefficient spectra reveals that the OAM modes with l = 25 is a dominant component (right panel). c The OAM weighting coefficient distributions for various injection currents within wthe orking range (left panel). Here, the pseudo-colors represent eight OAM modes with l = 22-29. After the enumeration algorithm, four encoding OAM modes that feature a distinguishing largest weighting coefficient have been selected (right panel). d The detailed illustration of the enumeration algorithm for the four current-addressed modes. First, the measured OAM weighting coefficients spectra are grouped, with each group containing four OAM mode components featuring distinct dominant OAM components. Second, the groups are evaluated to determine the one where the dominant OAM components show the greatest disparity in weighting coefficients compared to the other components within each spectrum. As a result, the four dominant OAM components from the selected group can be utilized to design the OAM-multiplexing hologram.

The design of the OAM-multiplexing hologram is illustrated in Fig. 4a (Supplementary Note 3). The Fourier holograms of four frames from a rotating windmill video are encoded by helical phase plates with OAM orders of l = −25, − 26, − 24 and − 23, then being superposed into the final OAM-multiplexing hologram. As a result, as the four selected injection currents (I = 28, 58, 67, and 101 mA) switched at fixed time intervals, the target holographic images with an average peak signal-to-noise ratio (PSNR) exceeding 18.8 dB can be dynamically reconstructed by selectively recovering the Gaussian mode pixels (Fig. 4b). In comparison to the low-quality images reconstructed by beams generated by the same VCSEL at the other randomly selected injection currents, the improvement of the average signal-to-noise ratio (SNR) suggests the effectiveness of the enumeration algorithm mentioned above (Supplementary Fig. 7). In addition, a spatial division scheme has been adopted in the design of the OAM-multiplexing hologram of a micro-quick response (MQR) code image, which leads to a current-shared holographic encryption. To construct the hologram, the MQR code image is divided into four sections, and sampled with a comb function (Fig. 4c). Then, the Fourier holograms are encoded with the selected helical phase plates (l = −25, − 26, − 24, and − 23), resulting in an OAM-multiplexing hologram. Therefore, after combining and binarizing the four reconstructive images obtained under the four selected injection currents, the complete MQR code image is obtained (Fig. 4d). Scanning the MQR code reveals the encrypted message “Hello World”.

Fig. 4: Current-addressed holographic video and current-shared holographic encryption scheme.
Fig. 4: Current-addressed holographic video and current-shared holographic encryption scheme.
Full size image

a The design process of the OAM-multiplexing hologram for the target video comprising four image frames. b The reconstructive video images (bottom row), when switching the appropriate injection currents (middle row) at a time sequence of 250 ms (upper row). c Design process of the OAM-multiplexing hologram of a MQR code image for current-shared holographic encryption. d Experimental reconstruction of the OAM-multiplexing hologram. After combining and binarizing the four reconstructive images, the complete MQR code image can be obtained. Scanning the MQR code reveals the encrypted message “Hello World”.

Ultra-compact holographic VCSEL chip with ultra-high frame rate

By directly printing OAM-multiplexing holograms onto the surface of the VCSEL (Methods, Supplementary Fig. 8), we successfully demonstrate a dynamic holographic VCSEL chip using the above-mentioned current-addressed modes multiplexing holography (Fig. 5). To experimentally characterize the performances of the chips, an on-chip measurement system has been developed (Fig. 5a), wherein the VCSEL chip was powered via a metal probe that directly contacted the p-electrode (inset in bottom right). When the currents were switched, independent holographic images can be observed in the charge-coupled device (CCD) camera (inset in upper right). In our scheme, four VCSELs with oxide aperture diameters of 8 μm, 10 μm, 12 μm, and 14 μm have been selected as the light sources to construct a device array (Fig. 5b and Supplementary Fig. 9). Considering the divergence angles of the VCSELs and the printed holder height (150 μm), the hologram size was set as 100 μm × 100 μm with a pixel size of 1 μm × 1 μm to ensure optimal illumination of each hologram (inset in bottom left). Notably, 8-level transmission-type phase modulation was adopted through controlling the height of each pixel pillar (inset in bottom right).

Fig. 5: Ultra-compact holographic VCSEL chip with ultra-high frame rate.
Fig. 5: Ultra-compact holographic VCSEL chip with ultra-high frame rate.
Full size image

a Photo of the on-chip measurement system to characterize the display performances (Left panel). The holographic VCSEL chip was powered via a metal probe that contacts the p-electrode (bottom right). The holographic images can be dynamically switched when changing the injection currents (upper right). b SEM images of a holographic VCSEL chip comprising 2*2 hologram devices array. c Current-switching holographic display. d Current-switching 3D holographic display at separate image planes. e The modulation response-frequency curve of the VCSEL chip, indicating a bandwidth of 1.93 GHz at −3 dB.

For holographic encoding in OAM multiplexing holography, two dominant OAM states at specific injection currents have been employed for each holographic VCSEL chip, respectively. Consequently, by switching the injection currents on each VCSEL (from 9 mA, 16 mA, 23 mA, 23 mA to 19 mA, 9 mA, 27 mA, 32 mA), the dynamic holographic display (from “HOLO” to “GRAM”) can be obtained (Fig. 5c). Moreover, the current-addressed dynamic holographic display can be extended to 3D images by incorporating lens functions with varying focal lengths, thereby leveraging the depth information of the image along the light propagation direction (Supplementary Fig. 10). As such, when the holographic VCSEL chip was given as 10 mA or 18 mA, independent switchable images (Arabic number images “2, 0” or “0, 4”) can be observed at separate image planes at a propagation distance of z = 215 μm and 495 μm (Fig. 5d and Supplementary Fig. 11). Notably, the modulation response-frequency curve of the VCSEL chip illustrates that the refresh rate of the dynamic holographic display can reach up to ~1.93 GHz (Fig. 5e and Supplementary Fig. 12, “Methods”). We also conducted standard measurements of the frequency-dependent properties of the used VCSEL (Supplementary Figs. 13–15, Methods), which show that its bandwidth reaches up to 3 GHz. The gap between this bandwidth and the observed holographic refresh rate is mainly due to differences in the testing methods. These results indicate that further improvements in the packaging of the VCSEL holographic chip may support even higher refresh rates. Additionally, a diffraction grating hologram was printed onto the surface of the VCSEL to evaluate the diffraction efficiency of such an optical digitalized hologram. The diffraction efficiency is quantified by the ratio of the intensity of the + 1-order diffraction spot to the total output intensity of the VCSEL. The result indicates the final average modulation efficiency of 44.68%, which is determined by the grating diffraction efficiency and the transmission efficiency of the substrate (Supplementary Fig. 16).

Discussion

In summary, we develop a holographic VCSEL chip that eliminates the barriers to achieving an ultra-compact holographic system with dynamic light control capability. In this context, the injection current-dependent dominant OAM component states in multimode VCSELs are revealed to achieve current-addressed mode-multiplexed holography. Through laser nanoprinting, OAM-multiplexing hologram (arrays) can be incorporated into the VCSELs, with each unit occupying merely hundreds of microns. A dynamic 3D holographic display is achieved by current-switching at frame rates up to ~ 1.93 GHz, enabled by the ultra-high modulation bandwidth of oxide-confined VCSEL architectures. Notably, VCSELs with larger oxide apertures support more spatial modes suitable for holographic encoding, thereby enhancing the multiplexing capacity (Supplementary Fig. S17). In addition, the principle of current-addressed mode-multiplexed holography can be applied in the design of metasurface holograms with subwavelength pixel resolution37,38, where polarization-multiplexing strategies may be incorporated to further increase the available holographic channel numbers39. Furthermore, VCSEL configurations, such as chaotic-cavity VCSELs40,41,42, warrant further exploration. Their irregular geometries support a substantially larger number of transverse modes, theoretically enriching the available channels for multiplexed hologram design. However, chaotic-cavity VCSELs suffer from mode complexity and instability, which can degrade holographic performance and stability. Therefore, further co-design and optimization of the VCSEL structure and hologram algorithm may be required to address these challenges. It can be expected that the experimental realization of such an integrated, cost-effective and dynamic holographic system opens avenues for transformative applications in wearable devices43, ultra-high-speed short-reach interconnects44, virtual and augmented reality45, and autonomous driving46. This work bridges nanophotonics and optoelectronics, paving the way for next-generation dynamic holography in both scientific and industrial domains.

Methods

COMSOL multiphysics simulation

We used COMSOL Multiphysics 6.2 to calculate the current density distribution shown in Fig. 2c. First, we defined the global parameters, including both structural and material parameters, based on the epitaxial structure of the VCSELs (Supplementary Table 1). Using rotational symmetry modeling, we constructed the geometry of the VCSEL and assigned the appropriate materials to each region. The VCSEL diameter was set as the oxide aperture diameter plus 11 μm, and the inner diameter of the annular p-contact was set as the oxide aperture diameter plus 3 μm, following the device specifications. Next, we discretized the structure to enable an efficient and accurate solution of the physical field. A direct bandgap spontaneous emission light transition model was applied, and the steady-state solutions were obtained through voltage sweeps and semiconductor equilibration. Finally, we analyzed the current distribution by examining a two-dimensional cross-section, located 50 nm above the active region, under different voltage conditions.

Optical setup for the measurement of OAM components weighting spectra of VCSELs and the current-addressed holographic video

To measure the OAM components weighting spectra of VCSELs and obtain the current-addressed holographic video, the conceptual illustration and the practical optical setup can be found in Fig. 3a and Supplementary Fig. 4, respectively. The VCSEL operates under continuous-wave conditions. After supplying power on the VCSEL, a × 10 objective lens was utilized to modulate the divergence angle of the light beams. Then, a half-wave plate in combination with a polarization beam splitter was employed to control the intensity of horizontal-polarized light beams continuously. After that, the beams illuminated on the spatial light modulator (SLM, model: PLUTO-NIR-010-A) and a CCD camera (model: Basler, acA2040-90uc) was put on the focal plane to capture the reconstructed holographic images. The OAM-multiplexing hologram in Fig. 4a and the helical phase plates with different OAM orders were imprinted on the SLM, respectively.

Laser nanoprinting of the holograms on the surfaces of the multimode VCSELs

A commercial photolithography system (Photonic Professional GT, Nanoscribe GmbH), equipped with a 780 nm, 80 MHz femtosecond laser, was used for printing. The process was conducted in a dip-in galvo-scan-mode configuration, utilizing the two-photon resist IP-Dip and a high numerical aperture objective lens (Plan-Apochromat 63 x/1.40 Oil DIC, Zeiss). The printed structure consists of three key components: four supporting pillars (20 μm × 20 μm × 140 μm each), a supporting table (120 μm × 120 μm × 10 μm), and the holographic plates (100 μm × 100 μm).

The specific structural parameters are listed below. For the pillars, the slicing distance (step size in the longitudinal direction) was set to 0.5 μm, and the hatching distance (step size in the transverse direction) was set to 0.2 μm to optimize printing speed. For the table, the slicing distance remained at 0.5 μm, but the hatching distance was reduced to 0.05 μm to ensure a flat surface, as the laser beam must pass through it. The hologram height was 150 μm, and the slicing distance for the holograms was set to 0.226 μm, corresponding to a phase modulation of π/4 for the 905 nm VCSELs, with a hatching distance of 0.05 μm.

The detailed printing process is illustrated in Supplementary Fig. 8. The optimized printing parameters were as follows: 52 mW laser power and 20000 μm/s scanning speed for the pillars, 46 mW and 20000 μm/s for the table, and 40 mW and 10000 μm/s for the holograms. After printing, the samples were developed by immersion in propylene glycol 1-monomethyl ether 2-acetate for 30 min, followed by rinsing in isopropanol for 5 min. The devices were then air-dried through evaporation.

Characterization of the holographic VCSEL chip

As shown in Supplementary Fig. 12, the frequency response of the VCSEL holographic chip was characterized using a vector network analyzer (Rohde & Schwarz ZNLE14). Here, VCSEL_2 in Fig. 5c was selected for testing and connected to the PCB via gold wire bonding, with external linkage to the driver power supply. First, harmonic modulation signals from the network analyzer were used to cyclically switch the injection current of VCSEL_2 between 9 mA and 16 mA, inducing periodic variations in the energy of each pixel featuring a Gaussian mode in the reconstructed ‘O’ pattern. Afterwards, the optical output from one pixel was coupled into a single-mode fiber and detected by a photodetector. The single-mode fiber ensured that the photodetector collected only the energy variations of the Gaussian mode within the pixel. Finally, the photodetector output was analyzed using the network analyzer to obtain the modulation response–frequency curve.

In addition, the frequency-dependent characteristics of the used VCSELs were independently measured by a professional VCSEL manufacturer (Berxel Photonics Co., Ltd.). For these tests, the VCSELs were bonded onto a substrate, and the current signal was applied via SG probes. Modulation-response–frequency curves were acquired using the same procedure as in the holographic refresh-rate measurement, except that the photodetector directly collected the total optical output of the VCSELs. For the spectral-versus-frequency characterization, an arbitrary waveform generator (Anritsu MP1900A) replaced the vector network analyzer to provide high-speed modulation, and the resulting optical signal was coupled into an optical spectrum analyzer (YOKOGAWA AQ6370D) for spectral analysis.

The fundamental DC characteristics of the VCSELs used in our experiments, including I–V curves, emission spectra, and device lifetime, were also measured at Berxel Photonics Co., Ltd. Device lifetime was evaluated using a TO-LD drawer aging resistance testing system. Specifically, 29 VCSELs with 10-µm oxide aperture diameters were tested at an injection current of 16 mA and a temperature of 85 °C for a total duration of 168 h. A typical failure criterion for VCSELs is defined as a 20% reduction in output power. All tested devices remained fully functional throughout the 168 h stress period, indicating a lifetime exceeding 168 h under accelerated conditions. Furthermore, assuming device operation at 50 °C under a drive current of 16 mA, and employing the established activation energy (Ea = – 1.3 eV) reported in prior studies47, together with the standard acceleration factor (AF) calculation model48, we obtain an AF of 95.81 for the conditions of 16 mA at 85 °C (aging) versus 16 mA at 50 °C (usage). Accordingly, the estimated lifetime of the VCSELs under 50 °C and 16 mA operating conditions exceeds 16,096 h.