Cavityless vertical linear optical amplifier

Abstract

The application discloses a cavity-free vertical linear optical amplifier, and relates to the fields of semiconductor optoelectronic devices and optical communication technology. The amplifier comprises an InP-based semiconductor layer stack, at least three gain units arranged in sequence along the normal direction of the epitaxial layer and electrically connected in series through a tunneling interconnection structure, a common optical channel overlapping with the current injection region of each active region, a micro-optical coupling structure, and an electrode structure. A pair of resonant enhancement cavity reflection structures are not arranged between the input optical interface and the output optical interface of the device, and the optical signal passes through each active region in a single pass along the normal direction of the epitaxial layer. The device is suitable for amplitude and phase fidelity amplification of high-speed coherent optical signals and intensity modulation direct detection optical signals, and is convenient for wafer-level testing and two-dimensional array integration.

Description

Technical Field

[0001] This invention relates to the fields of semiconductor optoelectronic devices and optical communication technology, and in particular to cavity-less vertical linear optical amplifiers. Background Technology

[0002] Semiconductor optical amplifiers are widely used in optical link power compensation, receiver preamplification, transmitter boost amplification, optical interconnection, and inter-chip and inter-board optical coupling. Traditional traveling-wave semiconductor optical amplifiers propagate optical signals along the in-plane direction of the epitaxial layer, achieving a long gain length. However, their input and output ports are located on the side of the chip, posing significant engineering challenges in terms of end-face processing quality, array coupling, wafer-level testing, and co-packaging with surface-emitting optical devices.

[0003] Vertical-cavity semiconductor optical amplifiers (VCAs) employ pairs of mirrors on either side of the gain region, causing the optical signal to travel back and forth multiple times within the resonant cavity to achieve enhanced gain. The gain, phase, and group delay response of this type of structure are strongly dependent on the cavity mode position, mirror reflectivity, and cavity length. Temperature drift, epitaxial manufacturing deviations, or input wavelength shifts can all cause the operating signal to become detuned relative to the resonant peak. For high-speed optical signals carrying broadband modulation, narrowband resonance can also introduce significant amplitude slopes, abrupt phase changes, and group delay fluctuations across the signal spectrum, degrading signal quality.

[0004] From a signal processing perspective, high-speed coherent optical communication utilizes both amplitude and phase information of the optical field. Modulation formats such as QPSK, QAM, and orthogonal frequency division multiplexing are highly sensitive to the flatness of the amplifier's complex transfer function, polarization-dependent gain, and dynamic gain-phase coupling. Although intensity modulation direct detection formats such as NRZ and PAM do not directly detect the optical phase, inconsistent high and low level gain compression, pattern-dependent gain caused by carrier recovery, and edge broadening and ringing caused by resonant filtering all affect eye diagram opening, extinction ratio, and bit error rate performance. Therefore, optical amplifiers for high-speed signals must not only provide sufficient optical gain but also maintain stable amplitude and phase transfer relationships within a specified input power range and modulation bandwidth.

[0005] Existing amplification concepts with single-pass propagation along the epitaxial layer normal can improve the coupling conditions between wafer-level testing and two-dimensional arrays. However, the physical propagation length along the normal is constrained by the total thickness of the stack, resulting in limited single-pass gain. Simply increasing the thickness of the active material is limited by factors such as epitaxial strain accumulation, carrier transport, internal absorption, and thermal management. On the other hand, if heavily doped tunnel junctions are introduced between multiple active regions to achieve electrical series connection, the absorption of free carriers in the tunnel junctions and the coherent reflections at numerous interlayer interfaces may form gain ripples in the passband, or even constitute parasitic feedback channels that induce self-excited oscillations.

[0006] Therefore, there is a need for a semiconductor optical amplifier device structure that can effectively accumulate the single-pass gain of multiple gain units along the epitaxial layer normal, does not depend on paired high-reflectivity cavity mirrors, and can maintain linear complex transfer characteristics within a high-speed signal bandwidth. Summary of the Invention

[0007] To achieve the above objectives, the present invention provides a cavity-free vertical linear optical amplifier, characterized in that it comprises: InP-based semiconductor stack, wherein the InP-based semiconductor stack has an input optical interface and an output optical interface on opposite sides, respectively; At least three gain units are sequentially arranged along the epitaxial layer normal of the InP-based semiconductor stack. Each gain unit includes an active region that is lattice-matched or strain-matched with the InP crystal. Two adjacent gain units are electrically connected in series through a tunneling interconnect structure. Each active region has a current injection region that overlaps with a common optical channel. A micro-optical coupling structure is formed in at least one of the input optical interface and the output optical interface, or is fixedly optically coupled to the at least one of them. The micro-optical coupling structure includes at least one of an input micro-optical coupling structure and an output micro-optical coupling structure, and is used to allow an input optical signal to sequentially pass through each of the active regions along the common optical channel. An electrode structure, including an input electrode and an output electrode, is used to provide a series bias current to each of the gain units; Wherein, no paired reflection structures for forming a resonant enhancement cavity are provided between the input optical interface and the output optical interface, so that the input optical signal passes through each of the gain units in a single path along the epitaxial layer normal and is then output by the output optical interface; The region between two adjacent active regions is called the interstage region. The optical thickness of the multiple interstage regions is pre-set to be non-equivalent, so that at least two of the interstage regions give different round-trip propagation phases to the internal reflected light within a predetermined working band, and the multiple interstage regions do not constitute a fixed periodic stack for coherent enhanced reflection. The cavity-free vertical linear optical amplifier is configured to operate in a linear gain operating region below the parasitic self-excited oscillation threshold, so that the output optical signal retains at least one of the amplitude modulation information and phase modulation information of the input optical signal.

[0008] The beneficial effects of this invention are: Compared to vertical-cavity optical amplifiers that rely on high-reflectivity mirrors at both ends for resonance enhancement, this invention allows the optical signal to pass through multiple electrically cascaded gain units in a single path along the epitaxial layer normal. The gain is accumulated by each active region along the single-path optical path, independent of the Fabry-Pérot resonance peak. Double-sided antireflection, non-level optical thickness, and return-mode mismatch collectively reduce effective feedback within the chip and package, which helps to broaden the usable wavelength and temperature range and reduce the risk of parasitic self-oscillation.

[0009] This invention avoids introducing photon storage and cavity detuning selection effects into the high-quality factor cavity, which helps reduce narrowband gain ripple, phase abruptness, group delay peaks, edge broadening, and ringing caused by cavity resonance. For high-speed coherent optical signals such as QPSK and QAM, a relatively flat complex transfer function helps maintain the amplitude ratio and phase relationship between in-phase and quadrature components, reducing constellation non-uniform stretching and frequency selectivity error vector amplitude degradation caused by cavity detuning.

[0010] For high-speed intensity modulation direct detection optical signals such as NRZ and PAM, when sufficient saturation margin is maintained at each stage, interstage capacity configuration and linear operating point control are beneficial to reduce high-level compression, reduce extinction ratio, reduce code-related gain, and improve eye diagram closure.

[0011] Micro-optical coupling structures match the beam, common current aperture, and active region, reducing active area and drive current while maintaining local light intensity and linear margin. This supports wafer-level testing, two-dimensional arrays, and vertical integration with surface-emitting devices, detectors, or fiber arrays. The near-circular aperture and near-rotational-symmetric optical structure also provide design freedom to reduce polarization-dependent gain, making them suitable for dual-polarization coherent links.

[0012] The two types of tunneling interconnect schemes allow for the selection of low-absorption planar interconnects within a common optical channel or peripheral interconnects bypassing the high-intensity central region, depending on absorption, resistance, and process complexity. The non-equivalent optical thickness and downstream saturation capability configurations are optimized for coherent reflection and stepwise saturation, respectively, thus enabling single-pass gain accumulation, broadband linearity, and thermal stability without relying on a resonant cavity. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in this invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0014] Figure 1 This is a schematic cross-sectional view of the cavity-free vertical linear optical amplifier of the present invention.

[0015] Figure 2This is a schematic diagram comparing a low-absorption planar tunneling junction with a peripheral tunneling interconnect structure having a central transparent optical channel.

[0016] Figure 3 This is a schematic diagram illustrating the principle of partial cancellation of the non-equivalent optical thickness inter-level region and the internal reflective phasor.

[0017] Figure 4 This is a schematic diagram showing the matching relationship between the beam waist, cascaded gain stack, and current injection region formed by the micro-optical coupling structure.

[0018] Figure 5 This is a schematic diagram of a low-feedback single-pass optical path and the linear gain operating region.

[0019] Figure 6 This is a schematic diagram of an amplification module and closed-loop control for high-speed coherent optical signals and intensity-modulated direct detection optical signals.

[0020] Appendix Label Reference Table: Detailed Implementation

[0021] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. It should also be noted that, to make the embodiments more comprehensive, the following embodiments are the best and preferred embodiments, and those skilled in the art can use other alternative methods to implement some well-known technologies; moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0022] For ease of understanding, the terms used in this instruction manual are defined as follows: The term "cavity-free" as used in this specification refers to a device that does not employ paired distributed Bragg mirrors, metallic high-reflectivity mirrors, or equivalent high-reflectivity structures for the purpose of creating resonance enhancement and amplification. Instead, it utilizes optical interface anti-reflection, interstage phase configuration, and optional return-mode mismatch to ensure that the round-trip mode gain recoupled to the common optical channel within the chip and package is below the self-oscillation threshold. "Cavity-free" does not require all material interfaces to have strictly zero Fresnel reflection, nor does it exclude the use of low-reflectivity functional layers for monitoring, stray light filtering, or mechanical protection purposes.

[0023] The term "linear optical amplification" as used in this specification refers to a device in which, within its operating domain defined by a predetermined signal wavelength, input optical power, polarization state, temperature, bias, and modulation bandwidth, the complex envelope of the output optical signal is approximately proportional to the complex envelope of the input optical signal. This linearity does not require the device to maintain strict mathematical linearity at arbitrary input power; the device does not rely on deep gain saturation, cross-gain modulation, cross-phase modulation, optical bistable mode, or cavity resonance as its basic amplification mechanism; fixed complex gain and fixed propagation delay, which can be calibrated or equalized at the receiver, are not considered to violate the linearity.

[0024] In this specification, "epitaxy layer normal" refers to the direction perpendicular to each epitaxial layer of the InP-based semiconductor stack, i.e., the epitaxial growth direction; "common optical path" refers to the common optical path region where the input optical signal passes through each active region sequentially along the epitaxial layer normal; "interstage region" refers to the region between the reference planes of two adjacent active regions, including tunneling interconnect structures and corresponding spacer layers; and "current injection region" refers to the effective injection region within each gain unit that is defined by a current-limiting structure and optically overlaps with the common optical path.

[0025] The term "parasitic self-oscillation threshold" as used in this specification refers to the round-trip mode gain of 1 when the returned light from the device and package is recoupled to the pumped common optical channel after one round trip (i.e., the return-to-round mode gain is equal to 1). The critical condition of ); the so-called "linear gain working region" refers to the bias-power-temperature working region below the parasitic self-excited oscillation threshold and where the gain compression of each level does not exceed the preset linearity upper limit; the so-called "equivalent saturated output capability" refers to the level of saturated output power that the gain unit can provide under given bias conditions, which is jointly determined by the active material, the number of quantum wells, the injection area and the mode field area.

[0026] Cascaded single-pass gain principle: Along the epitaxial layer normal, the physical gain length of a single gain cell is limited by the epitaxial layer thickness. Let the number of gain cells be N ( The modulus gain coefficient of the i-th gain unit is (in As the light limiting factor, (where is the material gain coefficient), and the active region gain length is... Then the single-pass power gain of the device can be expressed as: As can be seen, the gain is accumulated exponentially by each gain unit along a single optical path, without relying on multiple round trips to enhance the Fabry-Pérot resonant peak. Adjacent gain units are electrically connected in series through a tunneling interconnect structure, and the same series bias current is injected into each gain unit sequentially. The device terminal voltage is the sum of the junction voltage of each unit and the tunnel junction voltage drop, thereby realizing multi-stage gain regeneration under single-port drive.

[0027] Interstage non-uniform optical thickness configuration: The region between the reference planes of two adjacent active regions is designated as the i-th interstage region, which is designed for the vacuum wavelength. single-pass phase thickness at the location Represented as: in, and The m-th layer of material in the i-th interstage zone is respectively The refractive index and physical thickness at that location This refers to the optical thickness of the inter-level region. The optical thicknesses of multiple inter-level regions are arranged in a non-equivalent manner so that the reflected light from different internal interfaces does not accumulate in phase according to a single fixed period. The non-equivalent arrangement can be a linear chirped sequence, a nonlinear chirped sequence, an alternating sequence with three or more different optical thickness values, a quasi-periodic sequence, a quasi-random sequence, or an optimized sequence obtained by the transfer matrix method, the scattering matrix method, and numerical optimization.

[0028] Under the first-order interface reflection approximation, the complex amplitude reflection of the k-th principal internal interface can be expressed as: ,in The complex amplitude reflection coefficient of this interface. The reference plane represents the one-way propagation phase from the interface. By adjusting the interstage layer thickness, material composition, phase matching layers on both sides of the tunnel interconnect structure, and the interface position, multiple reflection phasors can be at least partially canceled out within a predetermined operating band. For structures with large refractive index differences or where multiple reflections are not negligible, the maximum effective reflection, gain ripple, and group delay ripple within the operating bandwidth are jointly optimized using the transmission matrix, scattering matrix, or time-domain electromagnetic simulation.

[0029] Parasitic feedback inhibition conditions: The input and output optical interfaces preferably each feature a broadband anti-reflection structure. At least one optical interface may also incorporate a locally tilted light-transmitting surface, a wedge-shaped transparent layer, an off-axis microlens, a non-rotationally symmetric freeform surface, or a stray light absorption region. This ensures that forward light still passes sequentially through each current injection region, while the residual reflected light exhibits angular, lateral, focal, or wavefront mismatch relative to the forward light. The effective mode power reflectivity observed from the cascaded gain stack towards the input and output ends is denoted as... and The effective power gain experienced during forward and reverse propagation is denoted as follows: and The effective power mode coupling factor between the returned light and the pumped common optical channel is denoted as . Then the parasitic round-trip power factor F is expressed as: The device satisfies the above conditions within its predetermined operating domain. When the forward and reverse effective power gains are the same... and Take the same value. This condition constrains the effective feedback of the entire chip and package, not just the reflectivity of the exposed interfaces at both ends; F<1 indicates that the device is below the parasitic self-excited oscillation threshold.

[0030] Micro-optical coupling and beam-aperture matching: Micro-optical coupling structures can be directly fabricated on the input-side substrate, output-side semiconductor surface, or transparent carrier layer, or optically coupled to corresponding optical interfaces through bonding, mounting, or wafer-level packaging. Their forms can include spherical or aspherical refractive microlenses, Fresnel structures, diffractive optical elements, gradient refractive index elements, metasurface optical elements, freeform surface optical elements, or combinations thereof. The beam waist is preferably located within or near the center of the cascaded gain stack, ensuring that the Rayleigh length covers each active region, thereby maintaining overlap between the beam spot at each active region and the corresponding current injection region. Rayleigh length With waist radius The relationship is: in, The operating wavelength is specified. A smaller optical waist is not always better; rather, it is determined by a combination of factors including allowable local light intensity, gain saturation margin, diffraction spread, absorption by tunneling interconnect structures, and coupling tolerance. By matching the beam and current aperture, the current injection area and parasitic capacitance can be reduced while satisfying linearity and saturation margin, facilitating two-dimensional array integration.

[0031] Tunneling interconnect structure: Tunneling interconnect structures can employ either low-absorption planar tunneling junctions located within a common optical channel, or peripheral tunneling structures bypassing the high-intensity region at the center of the common optical channel. These two types of structures can be used interchangeably between different stages of the same device. Low-absorption planar tunneling junctions reduce free carrier absorption and junction voltage drop by decreasing the optical thickness of heavily doped layers, adjusting bandgap alignment, selecting materials with larger band gaps, and employing graded doping or staggered heterostructures. Peripheral tunneling structures include continuous annular, segmented annular, or multiple arc-shaped peripheral tunneling regions surrounding a central transparent optical channel, as well as lateral current extension layers located on both sides of these regions. The driving current flows laterally to the peripheral tunneling regions, tunnels through, and then returns laterally to the next gain unit, while the main working light passes through the central transparent optical channel.

[0032] Linear gain operating range and operating point control: In the complex envelope representation, the device output can be written as: Where G is the dimensionless power gain. To adopt The transmission phase under the time factor convention, where t represents the slowly changing operating time variable relative to the optical carrier period. This indicates the amplification of spontaneous emission components. Within the linear gain operating region, H remains within the allowable error range under signal peak-valley power variations and predetermined environmental changes. The device's drive and monitoring controller can adjust the series bias current, device temperature, variable optical attenuation, or optical coupling state based on at least one of the following: input optical power, output optical power, reflected optical power, amplified spontaneous emission power, drive voltage, and device temperature. This prevents each gain unit from entering deep saturation and maintains parasitic feedback margin.

[0033] For high-speed optical signals occupying the signal bandwidth B, it can be used to The in-band power gain ripple is represented. Group delay ripple Evaluate the frequency domain linearity of a cavity-less amplifier: Cavity-free structures are used to reduce gain peaks, phase abrupt changes, and group delay peaks caused by resonant cavities; material gain spectra, residual interference of antireflection films, and carrier dynamics are optimized through active region design, interstage phase design, and operating point control. The dynamic gain changes caused by carriers are also affected by the linewidth enhancement factor. This is converted into a phase change; therefore, limiting the dynamic gain compression at each stage not only helps maintain the high-low level ratio of the intensity-modulated direct detection signal, but also helps reduce the amplitude-phase conversion of the coherent signal. For dual-polarization coherent signals, the current injection region, the central transparent optical channel, and the micro-optical elements can be set as near-circular or near-rotationally symmetric structures, and the polarization-related gain can be reduced through strain compensation and symmetrical electrodes; the residual fixed polarization difference can also be calibrated at the receiver.

[0034] Gain unit and material system: The active region of each gain unit can be constructed using InGaAsP, InAlGaAs, or AlGaInAs material systems that are lattice-matched or strain-matched to InP, forming multiple quantum wells, strain-compensated quantum well groups, short-period superlattices, bulk gain layers, quantum wires, or quantum dot structures. The device can cover one or more operating bands from the O, E, S, C, and L bands, with the predetermined operating band preferably located in... Range. The number of cascaded gain units can be set according to epitaxial strain, series voltage, heat dissipation capacity, target bandwidth, saturation margin, and parasitic feedback margin, for example, from six to twelve.

[0035] Each gain unit can have the same structure, or different active material thicknesses, quantum well numbers, current injection areas, beam areas, carrier injection capabilities, or equivalent saturation output capabilities can be set according to the optical power increasing step by step along the propagation direction. Preferably, the equivalent saturation capability of the downstream gain unit is not lower than that of the upstream gain unit to avoid concentrated depletion of carriers in a certain stage by optical power. A high thermal conductivity substrate, thermal diffuser, or flip-chip heat dissipation structure with a light-transmitting area can be provided on the output optical interface side to make the downstream gain unit bearing higher optical power close to the main heat dissipation path.

[0036] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the listed material systems, structural dimensions, signal rates, power levels, and linearity indicators are used to illustrate the design and verification methods and do not constitute a limitation on the scope of protection; when not explicitly stated as measured results, the relevant values ​​are optional design values ​​or test conditions. The technical features in each embodiment can be combined with each other as long as there are no contradictions, and the resulting technical solutions all fall within the scope of protection of the present invention. Specifically, Embodiment 1 provides the specific structure and fabrication process of the device; Embodiment 2 provides the linear amplification operation mode and linearity verification of high-speed coherent optical signals and intensity-modulated direct detection optical signals; Embodiment 3 provides the amplification module, closed-loop control scheme, and various structural modifications.

[0037] Overall structure and working principle of the device: like Figure 1 As shown, the cavity-less vertical linear optical amplifier 100 is based on an InP-based semiconductor stack 105. Along the normal of the epitaxial layer, the opposite sides of the InP-based semiconductor stack 105 form an input optical interface 101 and an output optical interface 130, respectively. An input anti-reflection structure 103 and an input micro-optical coupling structure 102 are sequentially arranged on one side of the input optical interface 101; an output anti-reflection structure 131 is arranged on one side of the output optical interface 130, and an output micro-optical coupling structure 132 and a thermally conductive substrate or thermal diffuser 140 with a light-transmitting region 141 can be provided as needed. The input electrode 104 and the output electrode 133 are positioned away from the common optical channel 115, and a series bias current is applied to the cascaded gain stack 110.

[0038] The cascaded gain stack 110 includes at least three gain units 111 arranged sequentially along the normal of the epitaxial layer. Each gain unit 111 includes an n-type InP injection layer, a lower confinement layer, an active region 112, an upper confinement layer, and a p-type InP injection layer. Adjacent gain units 111 are electrically connected in series through a tunneling interconnect structure 114. Because the gain units 111 are electrically connected in series, the same series bias current is injected sequentially into the active region 112 of each gain unit 111. This is equivalent to the same current being reused N times in the normal direction. Each stage of the active region 112 receives carrier injection and provides gain, thereby achieving multi-stage accumulation of gain in a single-path optical path in the normal direction.

[0039] The input optical signal 150 is coupled into the common optical channel 115 via the input micro-optical coupling structure 102. It passes through the current injection region 121 of each gain unit 111 sequentially along the epitaxial layer normal, gaining gain step by step to form the output optical signal 151, which is output from the output optical interface 130. The device does not form distributed Bragg mirrors or metal high-reflectivity mirrors on both sides of the cascaded gain stack 110 to form a resonant enhancement cavity. The optical signal passes through each active region 112 only once, so its gain response does not exhibit cavity mode selectivity. This avoids the problems of resonant peak detuning, narrowband gain ripple, abrupt phase change, and group delay peak from a mechanistic perspective.

[0040] Interstage non-uniform optical thickness design: like Figure 3 As shown, the interstage regions 113 between adjacent active regions 112 are given different optical thicknesses. The design does not simply place all active regions 112 at the same standing wave antinode, but rather selects an aperiodic phase thickness sequence with the common optimization objectives of maximizing effective reflection, gain ripple, group delay ripple, and manufacturing tolerance within the predetermined operating band: During the epitaxial design phase, the transfer matrix method is used to calculate the complex reflections at both optical interfaces and each internal interface, and the complex amplitude reflections of each major internal interface are then calculated. Phasor superposition is performed in the complex plane, and the refractive index of each layer in the inter-level region 113 is adjusted. With physical thickness This design ensures that the reflected phasors are at least partially destructive within the predetermined operating wavelength range. During the design process, tolerance optimization is performed for manufacturing thickness variations, material dispersion, and temperature changes to ensure that the interstage region 113 does not form a fixed-period coherent enhancement reflective layer stack within the range of process and operating conditions. This design supports both the low-absorption planar tunneling junction 117 within the common optical channel 115 and the peripheral tunneling interconnect structure, without requiring the tunneling junction to be located at a resonant standing wave node.

[0041] Two implementation schemes for tunnel interconnection structures: like Figure 2 As shown, the tunnel interconnection structure 114 can be implemented in two schemes, and the two schemes can be mixed and used between different levels of the same stack.

[0042] The first type of scheme is a low-absorption planar tunnel junction 117 located within the common optical channel 115. Its p++ and n++ layers are located within the common optical channel 115. By thinning the heavily doped layers to reduce their optical thickness, using graded doping, setting a wide bandgap transition layer with a bandgap higher than the working photon energy, or using an interleaved heterogeneous bandgap arrangement, the absorption of free carriers and junction voltage drop are reduced, so that the additional loss introduced by the tunnel junction to the forward light 152 is within an acceptable range.

[0043] The second approach involves an outer tunneling structure that bypasses the high-intensity central region. The outer tunneling region 116 surrounds the central transparent optical channel 118 and can be a continuous ring, a segmented arc-shaped region along the circumference, or multiple discrete tunneling regions. Lateral current extension layers 122 are connected to both sides of the outer tunneling region 116, extending from the outer tunneling region 116 to the corresponding current injection region 121. The driving current flows laterally through the lateral current extension layer 122 to the outer tunneling region 116, tunnels, and then returns laterally to the current injection region 121 of the next gain unit 111. The main working light propagates through the central transparent optical channel 118. The central transparent optical channel 118 is filled with lightly doped InP or a transparent semiconductor material with a band gap higher than the working photon energy, thereby avoiding absorption of the optical signal in the high-intensity region by the heavily doped tunneling material. The current injection region 121 of each gain unit 111 is defined by a current limiting structure 120, which can be formed by processes such as selective region epitaxy, etching backfill, ion implantation, oxidation, or insulating layer definition.

[0044] Micro-optical coupling and beam-current aperture matching: like Figure 4 As shown, the input micro-optical coupling structure 102 and / or the output micro-optical coupling structure 132 position the beam waist 154 of the input beam within or near the center of the cascaded gain stack 110, and according to... The determined Rayleigh length covers each active region 112, ensuring that the light spot at each active region 112 maintains optical overlap with the corresponding current injection region 121. The optical waist size is determined comprehensively based on the allowable local light intensity, gain saturation margin, diffraction spread, absorption of the tunnel interconnect structure 114, and coupling tolerance, rather than using the minimum light spot size as the sole design objective.

[0045] The beam radius at each active region 112 varies with its distance from the beam waist 154. The corresponding current injection regions 121 may have the same or different lateral dimensions to compensate for the size changes of the beam as it propagates along the cascaded gain stack 110, or to compensate for the differences in saturation characteristics of different gain units 111. By using near-circular current apertures, symmetrical electrodes, and near-rotationally symmetrical micro-optical elements, the structural differences experienced by the two orthogonal polarizations are reduced, thereby reducing polarization-dependent gain and adapting to dual-polarization coherent links.

[0046] Linear gain operating region and parasitic feedback suppression: like Figure 5As shown, the input optical signal 150 passes through each gain unit 111 only in the forward direction. The input anti-reflection structure 103 and the output anti-reflection structure 131 employ single-layer, multi-layer, graded refractive index, or subwavelength anti-reflection structures depending on the operating wavelength, polarization state, and incident angle. The non-uniform optical thickness of the interstage region 113 causes partial cancellation of internal reflection phasors. The tilted light-transmitting surface, wedge structure, off-axis optical axis, or non-rotationally symmetric surface shape causes the residual reflected light 153 to experience angular shift, lateral displacement, focal shift, or wavefront mismatch relative to the forward light 152. These three factors work together to ensure that the device and package meet the following requirements: That is, the residual reflected light 153 cannot form resonance enhancement, and the device operates below the parasitic self-oscillation threshold. The linear gain operating region is defined by the upper limit of gain compression, the output dynamic range, and the parasitic feedback margin; for example, the gain compression corresponding to the signal peak can be controlled to not exceed... Within the design scope, this value is only an example of linearity calibration and is not a limitation on the size of the device gain.

[0047] Example 1: Device Structure and Fabrication Process In this embodiment, the cascaded gain stack 110 includes six gain units 111 arranged sequentially along the epitaxial layer normal, and each active region 112 adopts an InAlGaAs / InP multi-quantum-well structure. In other schemes, the gain units 111 can be three or more of any epitaxially implementable number, and the active regions 112 can be replaced by InGaAsP / InP quantum wells, strain-compensated quantum well groups, short-period superlattices, bulk gain layers, quantum wires, or quantum dot gain regions.

[0048] The device is fabricated according to the following process steps: Gain units 111, tunnel interconnect structures 114, and interlayer layers of varying thicknesses are sequentially grown on an InP substrate via molecular beam epitaxy or metal-organic chemical vapor deposition; a current-limiting structure 120 and a current-injection region 121 are formed by selective region epitaxy, etching backfill, ion implantation, oxidation, or insulating layer definition; an input electrode 104, an output electrode 133, a passivation layer, and antireflection structures 103 and 131 are fabricated; a micro-optical coupling structure is formed by etching, laser direct writing, imprinting, bonding, or wafer-level mounting; finally, substrate thinning, flip-chip bonding, and heat dissipation packaging with a light-transmitting region 141 are completed. During epitaxy, the thickness of subsequent interlayer layers can be corrected based on in-situ measurements or optical thicknesses measured on a test piece to ensure the accuracy of the varying optical thickness sequence.

[0049] Example 2: Linear amplification of high-speed coherent and intensity-modulated direct detection signals This embodiment uses a working waveband located at Nearby InP-based devices are equipped with six to twelve InAlGaAs / InP or InGaAsP / InP multi-quantum-well gain units 111. Broadband anti-reflection structures 103 and 131 are set on the input and output optical surfaces, with non-uniform optical thicknesses in the inter-stage regions 113. The common current aperture is nearly circular, and input and / or output microlenses position the beam waist 154 within the cascaded gain stack 110. For progressively increasing optical power along the propagation direction, downstream gain units increase the number of quantum wells, carrier storage, current injection area, or effective mode field area to ensure each stage operates within a preset linear range. An output-side thermal diffuser 140 is used to reduce junction temperature variations in downstream gain units.

[0050] The device is driven by a DC series bias and operates within its calibrated input power and temperature range. The signal to be amplified is, for example, [signal name missing]. QPSK, 16QAM, or 64QAM coherent optical signals can also be The optical signal can be directly detected by NRZ or PAM4 intensity modulation. The rates mentioned above are for illustrative purposes only; the usable signal rate is determined by the material gain spectrum, amplified spontaneous emission, carrier recovery, residual gain and group delay ripple, optical coupling, and receiver bandwidth, and is not limited to the example values.

[0051] For coherent signals, the power gain ripple within the test signal bandwidth Group delay ripple The calculation of polarization-dependent gain, optical signal-to-noise ratio, error vector amplitude, and bit error rate is performed, and the constellation diagrams before and after amplification are compared. Fixed complex gain and fixed propagation delay can be removed before calculating the error vector amplitude to evaluate the additional errors caused by dynamic gain, dynamic phase, and frequency-selective response. The cavity-free structure eliminates amplitude-phase selectivity caused by cavity resonance; amplification of spontaneous emission and carrier-induced amplitude-phase transitions (longitude enhancement factor) is also considered. Coupling is still controlled through bias, input power, interstage saturation margin, and active region design.

[0052] The intensity-modulated signal is directly detected to test eye height, eye width, optical modulation amplitude, extinction ratio, TDECQ, and bit error rate. Carrier memory effects are checked using a code pattern containing long strings of "1", long strings of "0", and rapid level transitions. Cavity-free single-pass propagation avoids additional ringing caused by high-quality factor in-cavity energy storage and resonant filtering; maintaining saturation margin at each stage helps reduce high-level compression and code-related eye diagram closure. The device does not rely on carrier saturation or code transitions as its amplification mechanism.

[0053] Example 3: Amplification Module, Closed-Loop Control, and Structural Modification like Figure 6As shown, the amplification module including the cavity-less vertical linear optical amplifier 100 of the present invention further includes an input optical fiber 171, an input monitoring unit 161, an output monitoring unit 162, an output optical fiber 172, a drive and monitoring controller 160, and a temperature control unit 173. The input optical fiber 171 and the output optical fiber 172 can be coupled to a single device or a two-dimensional device array via microlens arrays, respectively. The monitoring unit can acquire the input optical power, output optical power, reflected optical power, or amplified spontaneous emission power through a low-reflection beam splitter structure, a peripheral sampling detector, or a transparent window on the electrode side.

[0054] The drive and monitoring controller 160 stores an operating domain map obtained from device calibration. This map correlates wavelength, total input power, peak-to-average power ratio, polarization state, temperature, and series bias with gain compression, output power, reflection margin, gain ripple, and group delay ripple. During operation, the controller adjusts at least one of the following based on the monitored values: series bias current, temperature control unit 173, input variable attenuator, and coupled microactuator, to maintain the device in the linear gain operating region. It also reduces the bias or input power when abnormal backlighting, overtemperature, or output compression is detected, thus achieving closed-loop protection.

[0055] For dual-polarization coherent signals, a polarization diversity optical path can be set at the input end, allowing the two orthogonal polarizations to enter two amplification units in the same array respectively; alternatively, a single near-circularly symmetrical common optical channel 115 can simultaneously amplify two orthogonal polarizations. For intensity-modulated direct detection array links, multiple amplifiers can be aligned in a two-dimensional array with multi-core optical fibers, fiber arrays, surface-emitting laser arrays, or detector arrays. The above modules can be used as receiver preamplifiers, transmitter boost amplifiers, loss compensators in optical switching nodes, or surface normal amplification units in free-space optical links.

[0056] In the structural variations, the input and output directions can be interchanged; the micro-optical coupling structure can be set on only one side or on both sides; the current injection region 121 can be circular, elliptical, polygonal, or a partitioned adjustable structure; the tunneling interconnect structure 114 can be entirely low-absorption planar structures, entirely peripheral structures, or a combination of both; each gain unit 111 can share a single series current, or bypasses, segmented electrodes, or fine-tuning branches can be provided to correct manufacturing differences. The above variations can be adopted as long as the optical signal passes through multiple electrically series gain units 111 in a single path along the epitaxial layer normal, the device does not rely on paired reflection structures to form resonance enhancement, and the predetermined amplitude or phase information is maintained through the interstage non-uniform optical thickness and linear operating configuration.

[0057] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0058] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A cavity-less vertical linear optical amplifier, characterized in that, include: InP-based semiconductor stack (105), the InP-based semiconductor stack (105) has an input optical interface (101) and an output optical interface (130) on opposite sides respectively. At least three gain units (111) are sequentially arranged along the epitaxial layer normal of the InP-based semiconductor stack (105). Each gain unit (111) includes an active region (112) that is lattice-matched or strain-matched with the InP crystal. Two adjacent gain units (111) are electrically connected in series through a tunnel interconnect structure (114). Each active region (112) has a current injection region (121) that overlaps with a common optical channel (115). A micro-optical coupling structure is formed in at least one of the input optical interface (101) and the output optical interface (130), or is fixedly optically coupled to the at least one. The micro-optical coupling structure includes at least one of an input micro-optical coupling structure (102) and an output micro-optical coupling structure (132), for causing the input optical signal (150) to pass sequentially through each of the active regions (112) along the common optical channel (115). as well as The electrode structure includes an input electrode (104) and an output electrode (133) for providing series bias current to each of the gain units (111); Wherein, no paired reflection structures for forming a resonant enhancement cavity are provided between the input optical interface (101) and the output optical interface (130), so that the input optical signal (150) passes through each of the gain units (111) in a single path along the epitaxial layer normal and is then output by the output optical interface (130). The region between two adjacent active regions (112) is designated as the interstage region (113). The optical thickness of the multiple interstage regions (113) is pre-set to be non-equivalent, so that at least two interstage regions (113) impart different round-trip propagation phases to the internal reflected light within a predetermined working band, and so that the multiple interstage regions (113) do not constitute a fixed periodic stack for coherent reflection enhancement. The cavity-free vertical linear optical amplifier (100) is configured to operate in a linear gain operating region below the parasitic self-excited oscillation threshold, such that the output optical signal (151) retains at least one of the amplitude modulation information and phase modulation information of the input optical signal (150).

2. The cavity-free vertical linear optical amplifier according to claim 1, characterized in that, The active region (112) includes at least one of InGaAsP / InP quantum well, InAlGaAs / InP quantum well, strain-compensated quantum well, short-period superlattice, bulk gain layer, quantum wire gain layer, and quantum dot gain layer, and the predetermined operating wavelength is located in Within the range.

3. The cavity-free vertical linear optical amplifier according to claim 1, characterized in that, The optical thickness of the multiple inter-level regions (113) is set according to a linear chirped sequence, a nonlinear chirped sequence, an alternating sequence with three or more different optical thickness values, a quasi-random sequence, a quasi-periodic sequence, or an optimized sequence obtained according to the internal complex reflection cancellation condition.

4. The cavity-free vertical linear optical amplifier according to claim 1, characterized in that, The input optical interface (101) and the output optical interface (130) are respectively provided with anti-reflection structures (103, 131). The anti-reflection structures (103, 131) and the multiple inter-stage regions (113) together ensure that the return light generated by the two end optical interfaces, the internal inter-layer interface and the packaged optical elements and recoupled to the common optical channel (115) has a round-trip mode gain lower than the self-excited oscillation threshold within the predetermined operating band, bias current, input optical power and temperature range.

5. The cavity-free vertical linear optical amplifier according to claim 1, characterized in that, The micro-optical coupling structure includes at least one of a refractive microlens, an aspherical microlens, a Fresnel structure, a diffractive optical element, a gradient refractive index element, a metasurface optical element, and a freeform surface optical element, and sets the beam waist (154) of the input beam within a cascaded gain stack (110) formed by the first active region (112) to the last active region (112), or at a position adjacent to the cascaded gain stack (110).

6. The cavity-free vertical linear optical amplifier according to claim 5, characterized in that, At least one of the input optical interface (101), the output optical interface (130), and the micro-optical coupling structure has an inclined light-transmitting surface, a wedge-shaped structure, an off-axis optical axis, or a non-rotationally symmetric surface shape, so that the forward light (152) maintains optical overlap with each of the current injection regions (121), and causes the residual return light (153) to have an angular offset, lateral displacement, focal offset, or wavefront mismatch relative to the forward light (152), so as to reduce the mode overlap between the residual return light (153) and the common optical channel (115).

7. The cavity-free vertical linear optical amplifier according to claim 1, characterized in that, The lateral dimension of each current injection region (121) is matched with the beam size at the plane where each gain unit (111) is located, and at least two current injection regions (121) have different lateral dimensions to compensate for the size change of the beam as it propagates along the cascaded gain stack (110) or the difference in saturation characteristics of different gain units (111).

8. The cavity-free vertical linear optical amplifier according to claim 1, characterized in that, At least one of the tunneling interconnect structures (114) is a low-absorption planar tunneling junction (117) located in the common optical channel (115). The low-absorption planar tunneling junction (117) reduces light absorption by at least one of the following: material composition, doping distribution and axial thickness. The low-absorption planar tunneling junction (117) reduces free carrier absorption and junction voltage drop by thinning the heavily doped layer, using graded doping, setting a wide bandgap transition layer with a bandgap higher than the working photon energy, or using an interleaved heteroband arrangement.

9. The cavity-free vertical linear optical amplifier according to claim 1, characterized in that, At least one of the tunneling interconnect structures (114) is a peripheral tunneling structure, which includes a continuous or segmented peripheral tunneling region (116) arranged around the central transparent optical channel (118), and a lateral current extension layer (122) located on both sides of the axial direction of the peripheral tunneling region (116) and extending from the peripheral tunneling region (116) to the corresponding current injection region (121); the central transparent optical channel (118) is filled with low-doped InP or a transparent semiconductor material with a band gap higher than the working photon energy, the driving current flows laterally through the lateral current extension layer (122) to the peripheral tunneling region (116), and returns laterally to the next gain unit (111) after tunneling, and the optical signal passes through the central transparent optical channel (118).

10. The cavity-free vertical linear optical amplifier according to claim 1, characterized in that, With the direction of the optical signal from the input optical interface (101) to the output optical interface (130) as the downstream direction, at least one of the active material thickness, number of quantum wells, injection area, beam area, carrier injection capability and equivalent saturation output capability of each gain unit (111) is non-decreasing along the downstream direction; a thermally conductive substrate or thermal diffuser (140) with a light-transmitting region (141) is provided on one side of the output optical interface (130).

Images