Bidirectional high-voltage-resistant ESD (Electro-Static Discharge) protection device and implementation method

Abstract

The invention discloses a bidirectional high-voltage-resistant ESD (Electro-Static Discharge) protection device and an implementation method, relates to the technical field of chips, and aims to improve the bidirectional high-voltage-resistant capability of the ESD protection device. The ESD protection device comprises a semiconductor substrate and a heterojunction structure formed on the semiconductor substrate, a protection module is arranged on the heterojunction structure, the heterojunction structure comprises a first semiconductor layer and a second semiconductor layer, and a two-dimensional electron gas 2DEG conductive channel is formed at an interface of the heterojunction structure; the first semiconductor layer and the second semiconductor layer are both of a multi-stage stacked structure, and the first metal electrode and the second metal electrode are symmetrically distributed on the two sides of the heterojunction structure and are in ohmic contact with the 2DEG channel respectively; the P-type doped regions are located on the two sides of the heterojunction structure, and the composite passivation layer covers the heterojunction structure and the doped regions. According to the invention, the bidirectional high-voltage resistance of the ESD protection device is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic chips, and more particularly to a bidirectional high-voltage-resistant ESD protection device and an implementation method thereof. Background Art

[0002] Bidirectional high-voltage-resistant ESD protection devices play a crucial role in various electronic devices, protecting circuits from electrostatic discharge (ESD) damage. Its working principle is that the bidirectional high-voltage-resistant ESD protection device can provide protection under both forward and reverse voltages. When an electrostatic pulse strikes, the voltage across its two ends rises rapidly. When the breakdown voltage of the device is reached, the device will conduct quickly, introducing the electrostatic current into the ground or a low-impedance path, thus preventing the electrostatic from damaging other sensitive components in the circuit.

[0003] Traditional ESD devices such as TVS diodes and varistors have insufficient protection ability under high voltage, slow response time, or high clamping voltage. The "bidirectional high-voltage resistance" mentioned by users means that the device needs to effectively suppress high-voltage ESD events in both forward and reverse directions and withstand higher voltages without being damaged. Existing technologies also use multi-layer structures or heterojunctions to improve the breakdown voltage resistance. For example, using a GaN / AlGaN heterojunction to generate a two-dimensional electron gas (2DEG) to enhance conductivity and breakdown voltage resistance. Bidirectional protection requires a symmetric structure design, such as P-type doped regions on both sides to form a reverse PN junction, so that the protection mechanism can be triggered in both positive and negative ESD events. The bidirectional high-voltage resistance ability of existing ESD protection devices lags behind. Summary of the Invention

[0004] Aiming at the deficiencies of the above technologies, the present invention discloses a bidirectional high-voltage-resistant ESD protection device and an implementation method thereof, using a composite passivation layer (alternating SiN and Al 2 O 3 alternately) to reduce leakage current and electric field concentration, optimizing the electric field distribution through gradient doping, improving the anti-breakdown ability of electrical energy by arranging an electrostatic discharge module on the semiconductor substrate, and improving the protection ability by arranging a protection module on the heterojunction structure on the semiconductor substrate.

[0005] The present invention adopts the following technical solutions: A bidirectional high-voltage-resistant ESD protection device, which includes: A semiconductor substrate, serving as the basic support structure of the entire device and providing a platform for the construction of other components; an electrostatic discharge module is arranged on the semiconductor substrate; A heterojunction structure formed on the semiconductor substrate, with a protection module disposed on the heterojunction structure. The heterojunction structure includes a first semiconductor layer and a second semiconductor layer, and a two-dimensional electron gas (2DEG) conductive channel is formed at the interface of the heterojunction structure. Both the first semiconductor layer and the second semiconductor layer are multi-stage stacked structures. The first semiconductor layer is gallium nitride (GaN), and the second semiconductor layer is aluminum gallium nitride (AlGaN). The polarization effect at the heterojunction interface forms a 2DEG channel, and the aluminum component ratio in the AlGaN layer is 25% - 40%. A first metal electrode and a second metal electrode symmetrically distributed on both sides of the heterojunction structure respectively form ohmic contacts with the 2DEG channel. P-type doping regions located on both sides of the heterojunction structure are formed by ion implantation or epitaxial growth and constitute a reverse PN junction with the 2DEG channel. The reverse PN junction is provided with a first P well composed of a first P+ heavily doped region and a first N+ heavily doped region, a second P well composed of a second N+ heavily doped region and a second P+ heavily doped region, and a third P well composed of a third P+ heavily doped region and a third N+ heavily doped region. The first P well, the second P well, and the third P well are arranged in parallel. The P-type doping region is provided with a strengthened metal field plate, and the field plate is connected to the source or the gate. The reverse PN junction is also provided with a field plate. A composite passivation layer covering the heterojunction structure and the doping region. The composite passivation layer includes at least two materials with different dielectric constants, which are used to suppress surface leakage and electric field concentration. The trigger voltage range is ±20V to ±20V, and the breakdown voltage withstand capacity is ≥60V.

[0006] As a further technical solution of the present invention, the electrostatic discharge module includes a first relay switch circuit and a second relay switch circuit connected in parallel. The first relay switch circuit is serially connected with a first transistor circuit and a second transistor circuit connected in series with each other. The second relay switch circuit is serially connected with a third transistor circuit and a fourth transistor circuit connected in series with each other. The second transistor circuit and the third transistor circuit are connected in parallel. The output end of the third transistor circuit is serially connected with a first current conversion circuit, and the output end of the fourth transistor circuit is serially connected with a second current conversion circuit.

[0007] As a further technical solution of the present invention, the protection module is a voltage sensing circuit based on a high-voltage turns ratio of 1:10. The output end of the voltage sensing circuit is connected to the input end of the diode protection circuit. The output end of the diode protection circuit is connected to the input end of the discharge circuit. The output end of the discharge circuit is connected to the input end of the output circuit.

[0008] As a further technical solution of the present invention, the composite passivation layer is made of a high-dielectric-constant material Al2 O 3 alternately deposited with a low dielectric constant material SiNx, where Al 2 O 3 The layer thickness is 5 - 20 nm, and the SiNx layer thickness is 50 - 200 nm; and the composite passivation layer includes a first passivation layer arranged side by side and a second passivation layer connected by a second transistor; a first transistor is arranged in the first passivation layer, and a third transistor is arranged in the second passivation layer; the first transistor is connected to the third transistor in a cascaded structure through the second transistor.

[0009] As a further technical solution of the present invention, the doping element of the P-type doping region is magnesium Mg, and the doping concentration is distributed in a gradient from the heterojunction interface to the surface, with a range of 1×10 18 cm −3 to 5×10 19 cm −3 .

[0010] As a further technical solution of the present invention, the first metal electrode and the second metal electrode form a charge emission region, the thickness of the first metal electrode and the second metal electrode is 610 - 800 nm, and it includes a stacked structure of titanium Ti, or aluminum Al, or nickel Ni, or gold Au.

[0011] The present invention also adopts the following technical solution: A method for realizing a bidirectional high-voltage-resistant ESD protection device, including: Step 1, preparing a semiconductor substrate, epitaxially growing a first semiconductor layer and a second semiconductor layer on the semiconductor substrate in sequence to form a heterojunction structure; the first semiconductor layer is gallium nitride GaN, the second semiconductor layer is aluminum gallium nitride AlGaN, and the aluminum component ratio of the AlGaN layer is 25% to 40%, and a two-dimensional electron gas 2DEG conduction channel is formed at the heterojunction structure interface; Step 2, setting a protection module on the heterojunction structure, depositing a metal layer on the heterojunction surface, forming a first metal electrode and a second metal electrode through an annealing process, and achieving ohmic contact with the 2DEG channel; P-type doping regions are formed on both sides of the heterojunction structure, formed by ion implantation or epitaxial growth, and form a reverse PN junction with the 2DEG channel; Step 3, a first P+ heavily doped region, a first N+ heavily doped region, a second N+ heavily doped region, a second P+ heavily doped region, a third P+ heavily doped region, and a third N+ heavily doped region are arranged in the reverse PN junction to form a first P well, a second P well, and a third P well, where the first P well, the second P well, and the third P well are arranged side by side; a strengthened metal field plate is arranged in the P-type doping region, and the field plate is connected to the source or the gate; a field plate is arranged on the reverse PN junction; Step 4: Cover a composite passivation layer over the heterojunction structure and the doping region. The composite passivation layer includes at least two materials with different dielectric constants and is used to suppress surface leakage current and electric field concentration. Form grooves on both sides of the heterojunction through photolithography and etching processes, and form a P-type doping region in the grooves by using an ion implantation process, with a doping concentration of 2×10 18 ∼6×10 19 cm −3 ; Adopt the atomic layer deposition (ALD) process to alternately deposit a high-dielectric-constant material Al 2 O 3 and a low-dielectric-constant material SiNx on the device surface to form a composite passivation layer.

[0012] As a further technical solution of the present invention, the epitaxially grown first semiconductor layer is GaN, and the second semiconductor layer is AlGaN; during application, adopt a pulsed doping process, and periodically inject silane SiH 4 gas during the growth of the AlGaN layer. The injection pulse frequency is 10∼50 Hz, and the pulse duty cycle is 20%∼80%, so that the AlGaN layer forms a longitudinal doping concentration gradient of 1×10 17 ∼5×10 18 cm −3 , and at the same time control the growth temperature at 1050±5 ∘ °C, and the V / III ratio is 2000∼3000 to increase the carrier mobility of the 2DEG channel to >2000 cm 2 / V⋅s. As a further technical solution of the present invention, the ion implantation of the P-type doping region adopts a dual-element co-doping process, with an implantation dose of 5×10 14 ∼2×10 15 cm −2 , where the atomic ratio of magnesium (Mg) to zinc (Zn) is controlled at 3:1∼5:1, and rapid thermal annealing is carried out in a nitrogen atmosphere at 800∼950 ∘ °C for 30∼120 s to form a gradient doping concentration distribution from 5×10 19 cm −3 at the surface to 1×10 18 cm −3 at the interface, and the reverse breakdown voltage is increased to >180 V; The annealing process of the metal electrode adopts a two-step method: The first stage: Anneal in a hydrogen atmosphere at 400∼500 °C for 5∼10 min to make the Ti / Al / Ni / Au stack react with the AlGaN surface to generate a TiN∼Al alloy transition layer; Second stage: Rapid annealing is carried out for 10 - 30 s in a nitrogen atmosphere at 750 - 850 °C to reduce the specific contact resistance to <5×10 −7 Ω·cm2, while suppressing the thermal decomposition of the AlGaN layer; During the ALD deposition of the composite passivation layer, every 5 - 10 nm of Al 2 O 3 is deposited, and then nitrogen plasma with 100 - 200 W is introduced for treatment for 10 - 30 s to reduce the density of interface states between layers to <1×10 11 cm −2 ·eV −1 , the surface leakage current is <0.5 nA, and the breakdown field strength is >15 MV / cm. As a further technical solution of the present invention, an improved sliding mode surface function is used to evaluate the leakage state when suppressing surface leakage, and the output of the improved sliding mode surface function is: In formula (1), represents the output of the sliding mode surface function, A represents the function stability parameter, represents the sliding mode surface function, represents the system state variable, u represents the continuity value of the sliding mode surface function, and b represents the state variable weight; where: In formula (2), i represents the number of leakage state types, n represents the total number of leakage state types, leakage state output function, represents the total set of leakage state output functions; the total set of functions is any one of the parameters of position, velocity, current or voltage; represents the state variable adjustment amount; the system state variable output function is: In formula (3), represents the state variable of the sliding mode variable structure control system; when calculating the system state variable output function, assume the sliding surface function Use the Lyapunov function to judge the stability of the sliding surface function, and the output of the judgment function is: Formula (4) is the expression of the Lyapunov function, and this function satisfies The function enters the sliding surface, and the system satisfies indicating that the surface leakage suppression effect is better; When discharging in the charge discharge area, the average values of active power and reactive power are as follows: In formula (5), is the average value of the active power output by the ESD protection device; is the average value of the reactive power output by the ESD protection device; is the voltage output by the first semiconductor layer (2) and the second semiconductor layer (3); is the sinusoidal current output by the first semiconductor layer (2) and the second semiconductor layer (3); The sliding mode surface function of the sliding mode variable structure outputs as: In Equation (6), are the sliding mode surface functions of the active power and reactive power of the electrostatic discharge module (101), respectively; is the preset value of the active power; P is the real-time value of the active power; is the preset value of the reactive power; Q is the real-time value of the reactive power; (I) is the leakage current output by the electrostatic discharge module (101), (V) is the voltage at the leakage point output by the electrostatic discharge module (101), and (Q) is the charge accumulation of the electrostatic discharge module (101). When the electrostatic discharge module (101) suppresses the surface leakage to the optimal state, the change rate of S is zero.

[0013] Positive and beneficial effects By providing an electrostatic discharge module on the semiconductor substrate, the present invention improves the safety ability, and by providing a protection module on the heterojunction structure, the high-voltage resistance ability is improved.

[0014] The present invention also adopts a bidirectional high-voltage resistance mechanism. A high-mobility 2DEG channel is formed through the polarization effect of the heterojunction (such as GaN / AlGaN), and the breakdown voltage can reach >1000V; the high-voltage resistance ability of the bidirectional high-voltage resistance mechanism is improved. The present invention forms a reverse PN junction between the symmetric P-type doping region and the 2DEG, triggering avalanche breakdown bidirectionally, and the clamping voltage V C <20V, improving the bidirectional high-voltage resistance and breakdown prevention ability.

[0015] The composite passivation layer of the present invention is formed by alternately depositing Al 2 O 3 (dielectric constant 9) and SiNx (dielectric constant 7), reducing the surface leakage current to <1nA and evenly distributing the electric field; The P-type doping region adopts a gradient doping process (concentration from 1×10 18 cm∼3 to 5×10 19 cm −3 ), optimizing the carrier distribution and enhancing the surge resistance ability. It has outstanding substantial features and significant technological progressiveness.

[0016] Through the synergistic effect of the heterojunction polarization effect, gradient doping, and composite passivation layer, this structure achieves bidirectional voltage resistance >600V and ESD protection ability >8kV.

[0017] The present invention improves the anti-breakdown ability of electric energy by providing an electrostatic discharge module on the semiconductor substrate, and improves the protection ability by providing a protection module on the heterojunction structure on the semiconductor substrate.

[0018] The process of the present invention also has the following technical advantages: Pulsed doping epitaxy: By dynamically modulating the doping gas flow rate, the longitudinal doping gradient of the AlGaN layer is achieved, and the contradiction between the 2DEG surface density and mobility is optimized.

[0019] Dual-element co-doping: Mg / Zn co-doping combined with gradient annealing improves the hole concentration and breakdown uniformity in the P-type region.

[0020] Two-step annealing electrode process: By controlling the temperature in stages, the contact resistance is reduced while avoiding thermal damage to AlGaN.

[0021] Plasma-assisted ALD: Interlayer nitrogen plasma treatment significantly reduces the interface state density of the passivation layer and improves the breakdown voltage reliability.

[0022] Figure 1 It is a schematic diagram of the principle of a bidirectional high-voltage-resistant ESD protection device of the present invention; Figure 2 It is a schematic diagram of the principle of the electrostatic discharge module in a bidirectional high-voltage-resistant ESD protection device of the present invention; Figure 3 It is a schematic diagram of the principle of the protection module in a bidirectional high-voltage-resistant ESD protection device of the present invention; Figure 4 It is a schematic diagram of the chip protection circuit in a bidirectional high-voltage-resistant ESD protection device of the present invention; Figure 5 It is a schematic diagram of an embodiment of anti-breakdown of a bidirectional high-voltage-resistant ESD protection device of the present invention; Figure 6 It is a schematic flow diagram of a method for realizing a bidirectional high-voltage-resistant ESD protection device; Reference numerals: 1 - Semiconductor substrate; 2 - First semiconductor layer; 3 - Second semiconductor layer; 4 - First P+ heavily doped region; 5 - First N+ heavily doped region; 6 - First P well; 7 - Second N+ heavily doped region; 8 - Second P+ heavily doped region; 9 - Second P well; 10 - Third P+ heavily doped region; 11 - Third N+ heavily doped region; 12 - Third P well; 131 - First metal electrode; 132 - Second metal electrode; 14 - First passivation layer; 15 - First transistor; 16 - Second transistor; 17 - Third transistor; 18 - Second passivation layer; 19 - First relay switch circuit; 20 - Second relay switch circuit, 21 - First transistor circuit; 22 - Second transistor circuit, 23 - Third transistor circuit; 24 - Fourth transistor circuit; 25 - First current conversion circuit; 26 - Second current conversion circuit; 101 - Electrostatic discharge module. Detailed implementation manners

[0023] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0024] As Figures 1-6 shown, a bidirectional high-voltage-resistant ESD protection device includes: A semiconductor substrate 1, serving as the basic support structure of the entire device and providing a platform for the construction of other components; an electrostatic discharge module 101 is disposed on the semiconductor substrate 1; A heterojunction structure formed on the semiconductor substrate 1, and a protection module is disposed on the heterojunction structure. The heterojunction structure includes a first semiconductor layer 2 and a second semiconductor layer 3. A two-dimensional electron gas 2DEG conductive channel is formed at the interface of the heterojunction structure; and both the first semiconductor layer 2 and the second semiconductor layer 3 are multi-level stacked structures; wherein the first semiconductor layer 2 is gallium nitride GaN, and the second semiconductor layer 3 is aluminum gallium nitride AlGaN. The polarization effect at the heterojunction interface forms a two-dimensional electron gas channel, and the aluminum component ratio of the AlGaN layer is 25% - 40%; A first metal electrode 131 and a second metal electrode 132 symmetrically distributed on both sides of the heterojunction structure, respectively forming ohmic contacts with the 2DEG channel; P-type doped regions on both sides of the heterojunction structure are formed by ion implantation or epitaxial growth and form a reverse PN junction with the 2DEG channel; the reverse PN junction is provided with a first P well 6 composed of a first P+ heavily doped region 4 and a first N+ heavily doped region 5, a second P well 9 composed of a second N+ heavily doped region 7 and a second P+ heavily doped region 8, and a third P well 12 composed of a third P+ heavily doped region 10 and a third N+ heavily doped region 11; wherein the first P well 6, the second P well 9, and the third P well 12 are arranged in parallel; the P-type doped region is provided with a reinforced metal field plate, and the field plate is connected to the source or the gate; the reverse PN junction is also provided with a field plate. A composite passivation layer covering the heterojunction structure and the doped region, the composite passivation layer comprising at least two materials with different dielectric constants, for suppressing surface leakage current and electric field concentration; the trigger voltage range is from ±20V to ±20V, and the breakdown voltage withstand capability is ≥60V.

[0025] In a specific embodiment, first, the electrostatic discharge module 101 identifies an abnormal situation by monitoring the voltage change across the device. When an electrostatic discharge pulse arrives, the voltage across the device will rise or fall rapidly. At this time, the 2DEG conductive channel formed by gallium nitride (GaN) and aluminum gallium nitride (AlGaN) in the heterojunction structure responds rapidly. The 2DEG channel has the characteristic of high carrier mobility, can efficiently conduct charges, and provides a low-impedance path for charges, thereby reducing the voltage drop and power consumption. On both sides of the heterojunction structure, the P-type doped region and the 2DEG channel form a reverse PN junction, which includes the first P well 6, the second P well 9, and the third P well 12. When an electrostatic discharge pulse arrives, the reverse PN junction will undergo avalanche breakdown. The avalanche breakdown is an electrical breakdown phenomenon. When the voltage exceeds a certain threshold, the electric field strength in the PN junction is large enough so that carriers can obtain sufficient energy to generate electron-hole pairs, resulting in a sharp increase in current. Through the good ohmic contact formed by the first metal electrode 131 and the second metal electrode 132 with the 2DEG channel, the electrostatic energy is guided to the ground terminal GND for discharge. Each P well is composed of a heavily doped P+ region and an N+ region, and multiple P wells are arranged in parallel to increase the breakdown voltage withstand capability and current-carrying capacity of the device; the P-type doped region is provided with a reinforced metal field plate, and the field plate is connected to the source or the gate. The metal field plate helps to optimize the electric field distribution, avoid excessive local electric field concentration, and thus prevent the device from being damaged due to excessive electric field strength. Secondly, the composite passivation layer covering the device surface is made of aluminum oxide (Al 2 O 3It is composed of silicon nitride SiNx with a low dielectric constant. The composite passivation layer utilizes the characteristics of materials with different dielectric constants to effectively adjust the surface electric field distribution, reduce the peak value of the electric field intensity, and suppress the occurrence of surface leakage current. Finally, the above-mentioned parts are integrated onto a semiconductor substrate to form a complete bidirectional high-voltage-resistant ESD protection device, ensuring the stability of the device under high voltage and the ESD protection performance, capable of achieving a breakdown voltage of not less than 60V, and providing an effective ESD protection function within the trigger voltage range of ±20V.

[0026] Further, the electrostatic discharge module 101 includes a first relay switch circuit 19 and a second relay switch circuit 20 connected in parallel. The first relay switch circuit 19 is serially connected with a first transistor circuit 21 and a second transistor circuit 22 connected in series with each other. The second relay switch circuit 20 is serially connected with a third transistor circuit 23 and a fourth transistor circuit 24 connected in series with each other. The second transistor circuit 22 and the third transistor circuit 23 are connected in parallel. The output end of the third transistor circuit 23 is serially connected with a first current conversion circuit 25, and the output end of the fourth transistor circuit 24 is serially connected with a second current conversion circuit 26.

[0027] In a specific embodiment, the first relay switch circuit 19 and the second relay switch circuit 20 switch the conduction state in real time based on the input voltage polarity. When the input voltage is a positive electrostatic pulse, the first relay switch circuit 19 identifies the positive polarity signal through a built-in voltage comparator, triggers the internal electromagnetic coil to act, closes the contact, and connects the first transistor circuit 21 and the second transistor circuit 22 to the main circuit. The control module outputs a driving voltage Vgsp according to the polarity signal and applies it to the gate of the first transistor circuit 21 to make it conduct quickly. The electrostatic current flows from the input end through the first transistor circuit 21 to the node N3. At this time, the second transistor adjusts its channel resistance through the bias voltage Vbc to efficiently direct the current to the output end Vout and avoid local voltage spikes. The cascaded design of the first transistor circuit and the second transistor can shorten the conduction time to <10ns to ensure rapid energy dissipation. The conduction current enters the first current conversion circuit 25, and the first current conversion circuit is composed of an LC resonance network and a transient voltage suppressor TVS. The LC network absorbs the high-frequency pulse energy through resonance, and the TVS clamps the remaining voltage to a safe value and finally discharges it to GND through a low-impedance path to complete the electrostatic discharge.

[0028] When the input is a negative electrostatic pulse, the second relay switch circuit 20 determines through the reverse voltage threshold, closes the contacts, and activates the third transistor circuit 23 and the fourth transistor circuit 24. At this time, the negative polarity trigger signal Vgsn acts on the gate of the fourth transistor circuit 24 to make it conduct. The current flows from GND through the third transistor circuit 23 and the node N2 in reverse conduction to the input end, forming a closed loop. The body diode characteristic of the third transistor circuit 23 further reduces the impedance during reverse conduction to prevent reverse breakdown. The second current conversion circuit 26 adopts a symmetric design, but introduces a reverse diode array. Utilizing its low forward voltage drop characteristic, it quickly guides the negative energy to GND, while avoiding reverse leakage current, and completes the electrostatic discharge.

[0029] Further, as Figure 3 shown, the protection module is a voltage sensing circuit based on a high-voltage turns ratio of 1:10; the output end of the voltage sensing circuit is connected to the input end of the diode protection circuit, the output end of the diode protection circuit is connected to the input end of the discharge circuit, and the output end of the discharge circuit is connected to the input end of the output circuit.

[0030] In a specific embodiment, in the Ethernet interface circuit, the 10 / 100 Ethernet PHY chip is the core and is responsible for the processing of the data link layer. The 10 / 100 Ethernet PHY chip generates transmission signals (such as TX+, TX-) and receives external input differential signals (such as RX+, RX-), and completes the link layer operations of data encoding and decoding. The Ethernet transformer transmits the differential signals through electromagnetic coupling, effectively isolates the DC component, suppresses the common-mode interference, and improves the anti-interference ability and integrity of signal transmission. At the same time, when the SLVU2-8 electrostatic discharge protection chip detects an electrostatic overvoltage, the internal path is quickly conducted, and the electrostatic energy is discharged to the ground (GND) to protect the backend circuit from electrostatic damage. The filter capacitor is connected in parallel between VCC and GND to filter out the high-frequency noise of the power supply and stabilize the supply voltage. The RJ45 interface is used as an external network connection port to transmit the processed signals such as TX-, RX- to external devices (such as network cables) to achieve physical layer communication. In addition, the voltage sensing circuit based on a high-voltage turns ratio of 1:10 monitors the transmission line voltage in real time, converts the high-voltage signal at a ratio of 1:10, and triggers the subsequent protection circuit when an abnormal overvoltage (such as a classic pulse) is recognized. The diode protection circuit conducts after receiving the signal of the voltage sensing circuit. Utilizing the diode clamping characteristic, it limits the voltage within a safe range to prevent the subsequent circuit from being subjected to overvoltage impact. The circuit also includes current-limiting resistors and impedance matching resistors (such as a 2.2Ω resistor to limit the circuit current), and the protection device is an impedance matching network composed of 75Ω resistors to optimize the signal transmission impedance and reduce the reflection loss. Further, the composite passivation layer is made of a high-dielectric constant material Al 2 O 3and an alternating deposition with a low dielectric constant material SiNx, where Al 2 O 3 The layer thickness is 5 - 20 nm, and the SiNx layer thickness is 50 - 200 nm; and the composite passivation layer includes a first passivation layer 14 arranged side by side and a second passivation layer 18 connected through a second transistor 16; a first transistor 15 is arranged in the first passivation layer 14, and a third transistor 17 is arranged in the second passivation layer 18; the first transistor 15 and the third transistor 17 are connected in a cascade structure through the second transistor 16.

[0031] In a specific embodiment, when the device is operating normally, Al with different dielectric constants 2 O 3 and the SiNx layer cooperate to adjust the surface electric field distribution and suppress surface leakage current. At this time, the cascaded transistors are in a cut-off or low-conduction state, finely adjusting the electric field and carrier distribution in the passivation layer. The Al 2 O 3 layer, due to its high dielectric constant, can effectively store charges, reduce the accumulation of surface charges, and thus reduce the surface leakage current. The SiNx layer, through its low dielectric constant, disperses the electric field, avoids the over-concentration of the electric field, and further suppresses leakage. The first transistor 15 and the third transistor 17, under the control of the second transistor 16, maintain a low conduction state, finely adjust the electric field and carrier distribution in the passivation layer, ensure the uniform distribution of the electric field on the device surface, and maintain the stable operation of the device.

[0032] When an abnormal overvoltage situation occurs, Al 2 O 3 and the SiNx layer respond quickly to reduce the peak value of the electric field intensity. The Al 2 O 3 layer can quickly capture and release charges, adjust the surface electric field, and reduce the sudden change of the electric field. The SiNx layer, through electric field dispersion, further reduces the peak value of the electric field intensity, preventing the device from breakdown due to excessive local electric field. The first transistor 15 detects the change in the electric field, quickly adjusts its conduction current, transmits and amplifies the signal through the second transistor 16, and enables the third transistor 17 to correspondingly adjust the electric field and carrier distribution in the second passivation layer 18. This synergistic effect can further suppress surface leakage and electric field concentration, protecting the device from electrostatic damage. After the abnormal overvoltage disappears, the electric field gradually returns to normal, and the transistors return to the initial stable state, continuing to finely adjust the situation in the passivation layer to ensure the stable operation of the device surface. Through the synergistic effect of the material properties of the composite passivation layer and the cascaded structure of the transistors, the entire process effectively controls the surface electric field and leakage of the device, improving the stability and reliability of the device in a high-voltage environment.

[0033] Further, the doping element of the P-type doping region is magnesium (Mg), and the doping concentration is distributed in a gradient from the heterojunction interface to the surface, with a range of 1×10 18 cm −3 to 5×10 19 cm −3 .

[0034] In a specific embodiment, in the P-type doping region, Mg has a low ionization energy (about 170 meV), and can be effectively ionized at room temperature to provide holes, improving the conductivity of the P-type doping region. The gradient doping design keeps the doping concentration at the heterojunction interface low at 1×10¹ 8 cm⁻³, reducing the electron scattering of the 2DEG channel, while the high surface doping concentration of 5×10¹ 9 cm⁻³ forms a low-resistance contact layer, enhancing the carrier injection efficiency of the PN junction. The atomic radius of the Mg atom has a high lattice matching degree with GaN, and the atomic radius difference is <5%, reducing the lattice defects introduced by doping, suppressing the proliferation of dislocations, improving the crystal quality of the material, and having good high-temperature stability, ensuring the stability of the doping distribution in subsequent processes. The space charge region formed by the gradient doping is in dynamic equilibrium with the electron gas of the 2DEG at the heterojunction interface. Combining with the multi-stage P-well structure, the peak value of the reverse electric field is reduced by about 30%. The high surface doping concentration region forms a depletion layer extension effect, making the electric field distribution uniform, increasing the breakdown voltage to >180 V, and at the same time suppressing the reverse leakage current <0.5 nA. The Mg ion implantation process is mature. By precisely controlling the implantation energy of 50 - 200 keV and the dose of 1×10¹ 4 -5×10¹ 5 cm⁻², the gradient doping from the heterojunction interface to the surface is realized, and combined with rapid thermal annealing to activate the doping, ensuring the consistency of device performance.

[0035] Further, the first metal electrode 131 and the second metal electrode 132 constitute a charge release region, and the thicknesses of the first metal electrode 131 and the second metal electrode 132 are 610 - 800 nm, and they include a stacked structure of titanium (Ti) or aluminum (Al) or nickel (Ni) or gold (Au).

[0036] In a specific embodiment, a multi-layer metal stacked structure of titanium (Ti) or aluminum (Al) or nickel (Ni) or gold (Au) is adopted, and the functions of each layer are synergistically complementary; the Ti titanium layer (5 - 10 nm) serves as an adhesion layer, and by forming a chemical bond with the GaN surface (such as a TiN interface layer), the bonding force between the electrode and the heterojunction is enhanced, preventing metal detachment; the Al aluminum layer (300 - 500 nm) serves as the main conducting layer, and due to the high conductivity of Al (electrical conductivity ≈ 3.5×10 7The Ni layer (50-100nm) serves as a diffusion barrier layer to inhibit the thermal diffusion of Al into the GaN layer and avoid the formation of harmful intermetallic compounds during high-temperature annealing (such as 750-850℃ nitrogen annealing); the Au layer (50-100nm) serves as a protective layer to prevent electrode oxidation and provide good welding performance. At the same time, the high chemical stability of Au (standard electrode potential +1.5V) ensures long-term reliability.

[0037] The total thickness of the metal stacked structure is 610-800 nm, of which Al accounts for 60-70%. At the same time, a two-step annealing process of 400-500°C hydrogen annealing and 750-850°C nitrogen annealing is combined to form Ti-Al alloy and NiAl intermetallic compound, achieving ultra-low contact resistance <5×10⁻ 7 Ω·cm² and stable interface. The electrode edge adopts a 15°-30° gradient bevel design, and the field plate structure is used to reduce the electric field concentration factor to below 1.5, inhibiting edge breakdown; the high work function of the Au layer (5.1eV) and the thickness control of the Ni layer (50-100 nm) optimize the Schottky barrier (0.8-1.2eV) and balance carrier injection and blocking.

[0038] The present invention also adopts the following technical solutions: Furthermore, a method for realizing a bidirectional high-voltage ESD protection device includes: Step 1: prepare a semiconductor substrate 1, and sequentially epitaxially grow a first semiconductor layer 2 and a second semiconductor layer 3 on the semiconductor substrate to form a heterojunction structure; the first semiconductor layer 2 is gallium nitride GaN, the second semiconductor layer 3 is aluminum gallium nitride AlGaN, and the aluminum component of the AlGaN layer accounts for 25% to 40%, and a two-dimensional electron gas 2DEG conductive channel is formed at the interface of the heterojunction structure; Step 2: a protection module is arranged on the heterojunction structure, and a metal layer is deposited on the surface of the heterojunction, and a first metal electrode 131 and a second metal electrode 132 are formed by an annealing process, and ohmic contact is achieved with the 2DEG channel; a P-type doped region is formed on both sides of the heterojunction structure, which is formed by ion implantation or epitaxial growth, and forms a reverse PN junction with the 2DEG channel; Step 3: a first P+ heavily doped region 4, a first N+ heavily doped region 5, a second N+ heavily doped region 7, a second P+ heavily doped region 8, a third P+ heavily doped region 10 and a third N+ heavily doped region 11 are arranged in the reverse PN junction to form a first P well, a second P well 9 and a third P well 12, wherein the first P well 6, the second P well 9 and the third P well 12 are arranged in parallel; an enhanced metal field plate is arranged in the P-type doped region, and the field plate is connected to the source or the gate; and an additional field plate is arranged on the reverse PN junction; Step 4: Cover a composite passivation layer over the heterojunction structure and the doping region. The composite passivation layer includes at least two materials with different dielectric constants and is used to suppress surface leakage current and electric field concentration. Form grooves on both sides of the heterojunction through photolithography and etching processes, and form a P-type doping region in the grooves by using an ion implantation process. The doping concentration is 2×10 18 ∼6×10 19 cm −3 ; Adopt the atomic layer deposition (ALD) process to alternately deposit a high-dielectric-constant material Al 2 O 3 and a low-dielectric-constant material SiNx on the device surface to form a composite passivation layer.

[0039] In a specific embodiment, first, select a semiconductor material such as silicon (Si) or silicon carbide (SiC) as the substrate; then epitaxially grow a first semiconductor layer of gallium nitride (GaN) and a second semiconductor layer of aluminum gallium nitride (AlGaN) on the semiconductor substrate in sequence to form a heterojunction structure. Control the aluminum component ratio of the AlGaN layer to be between 25% and 40% to ensure the formation of a 2DEG conductive channel at the interface of the heterojunction structure; design and set a protection module on the heterojunction structure to enhance the ESD protection ability of the device; deposit a metal layer on the surface of the heterojunction, and form a first metal electrode and a second metal electrode through an annealing process. The electrodes achieve good ohmic contact with the 2DEG channel; form a P-type doping region on both sides of the heterojunction structure by ion implantation or epitaxial growth, and make it form a reverse PN junction with the 2DEG channel; set a first P+-heavily doped region, a first N+-heavily doped region, a second N+-heavily doped region, a second P+-heavily doped region, a third P+-heavily doped region, and a third N+-heavily doped region in the reverse PN junction to form first, second, and third P wells, which are arranged in parallel to optimize the electric field distribution and breakdown voltage performance of the device; set a strengthened metal field plate in the P-type doping region and connect it to the source or gate to improve the electric field distribution of the device and increase the breakdown voltage ability; set a field plate on the reverse PN junction to further optimize the electric field distribution; cover a composite passivation layer including at least two materials with different dielectric constants over the heterojunction structure and the doping region to suppress surface leakage current and electric field concentration; form grooves on both sides of the heterojunction through photolithography and etching processes, and form a P-type doping region in the grooves by using an ion implantation process. Control the doping concentration to be between 2×10 18 ∼6×10 19 cm −3 ; finally, adopt the atomic layer deposition (ALD) process to alternately deposit a high-dielectric-constant material Al 2 O 3 and a low-dielectric-constant material SiNx on the device surface to form a composite passivation layer to further improve the insulation performance and breakdown voltage ability of the device.

[0040] Further, the first epitaxial semiconductor layer 2 is GaN, and the second semiconductor layer 3 is AlGaN; in the application process, a pulsed doping process is adopted, and silane SiH is periodically injected during the growth of the AlGaN layer 4 gas, the injection pulse frequency is 10∼50Hz, and the pulse duty cycle is 20%∼80%, so that the AlGaN layer forms a longitudinal doping concentration gradient of 1×10 17 ∼5×10 18 cm −3 ⁻³, and at the same time, the growth temperature is controlled at 1050±5 ∘ °C, the V / III ratio is 2000∼3000, so as to increase the carrier mobility of the 2DEG channel to >2000 cm 2 ² / V⋅s.

[0041] In a specific embodiment, a metal organic chemical vapor deposition (MOCVD) device is used to grow a gallium nitride (GaN) layer on a semiconductor substrate. During the growth process, the growth temperature, pressure, and gas flow rate are strictly controlled to ensure that the GaN layer has good crystal quality and the required thickness. An aluminum gallium nitride (AlGaN) layer is grown on the GaN layer. The aluminum component ratio is precisely controlled between 25% and 40%, the growth temperature is strictly controlled at 1050±5°C, and the V / III ratio is controlled between 2000 and 3000 to ensure the compositional uniformity and good crystal structure of the material. During the growth of the AlGaN layer, a pulsed doping process is adopted. By periodically injecting silane SiH 4 gas, a longitudinal doping concentration gradient is formed. In specific operations, the injection pulse frequency is set to 10 to 50Hz, and the pulse duty cycle is 20% to 80%, so that the doping concentration range in the AlGaN layer is between 1×10¹ 7 to 5×10¹ 8 cm⁻³. This gradient distribution helps to optimize the electric field distribution of the device, improve its breakdown voltage performance and ESD protection ability. During the growth process, advanced monitoring technologies are used to monitor the growth of the AlGaN layer in real time, including parameters such as thickness, composition, and doping concentration. According to the monitoring data, the growth conditions (such as temperature, pressure, and gas flow rate) are adjusted in a timely manner to ensure that the carrier mobility of the two-dimensional electron gas (2DEG) channel is increased to greater than 2000 cm² / V·s. After the growth of the AlGaN layer is completed, an annealing treatment is performed to eliminate the internal stress and crystal defects generated during the growth process, and further optimize the performance of the device. The parameters of the annealing temperature, time, and atmosphere are precisely controlled to effectively improve the crystal quality and electrical properties of the material. Further, the ion implantation of the P-type doping region adopts a dual-element co-doping process, and the implantation dose is 5×10 14 ∼2×10 15 cm −2, where the atomic ratio of magnesium (Mg) to zinc (Zn) is controlled to be 3:1 to 5:1, and rapid thermal annealing is carried out in a nitrogen atmosphere at 800 - 950 ∘ °C for 30 - 120 s, forming a gradient doping concentration distribution from the surface of 5×10 19 cm −3 to the interface of 1×10 18 cm −3 and increasing the reverse breakdown voltage to >180 V; The annealing process of the metal electrode adopts a two-step method: The first stage: Anneal in a hydrogen atmosphere at 400 - 500 °C for 5 - 10 min to make the Ti / Al / Ni / Au stack react with the AlGaN surface to form a TiN - Al alloy transition layer; The second stage: Rapidly anneal in a nitrogen atmosphere at 750 - 850 °C for 10 - 30 s to reduce the specific contact resistance to <5×10 −7 Ω·cm² and simultaneously inhibit the thermal decomposition of the AlGaN layer; During the ALD deposition of the composite passivation layer, every 5 - 10 nm of Al 2 O 3 is deposited, and then 100 - 200 W of nitrogen plasma is introduced for 10 - 30 s to reduce the interfacial state density between layers to <1×10 11 cm −2 ·eV −1 , with a surface leakage current of <0.5 nA and a breakdown field strength of >15 MV / cm.

[0042] In a specific embodiment, during the preparation process of the bidirectional high-voltage-resistant ESD protection device, the P-type doping region, the metal electrode, and the composite passivation layer are finely processed to improve the device performance. First, for the P-type doping region, a dual-element co-doping process is adopted, and magnesium (Mg) and zinc (Zn) are configured according to an atomic ratio of 3:1 - 5:1, and ion implantation is carried out with an implantation dose of 5×10¹ 4 -2×10¹ 5 cm⁻². Secondly, rapid thermal annealing is carried out in a nitrogen atmosphere at 800 - 950 °C for 30 - 120 s, which promotes the full diffusion and activation of the implanted ions and prevents surface oxidation. Finally, the annealing conditions are controlled to form a P-type doping region with a concentration ranging from 5×10¹ 9 cm⁻³ at the surface to 1×10¹ 8The gradient doping concentration distribution of cm⁻³ optimizes the electric field distribution, reduces the surface electric field concentration, and successfully increases the reverse breakdown voltage to >180V. The metal electrode adopts a Ti / Al / Ni / Au laminated structure, and its annealing process is divided into two steps. In the first stage, anneal in a hydrogen atmosphere at 400 - 500°C for 5 - 10 minutes. The hydrogen removes the oxide layer on the metal surface. At the same time, the metal laminate reacts with the AlGaN surface to form a TiN - Al alloy transition layer, enhancing the adhesion between the metal and the semiconductor. In the second stage, perform rapid annealing in a nitrogen atmosphere at 750 - 850°C for 10 - 30 seconds to further promote atomic diffusion and chemical reactions, reducing the specific contact resistance to <5×10⁻ 7 Ω⋅cm², and the nitrogen atmosphere inhibits the thermal decomposition of the AlGaN layer, ensuring the stability of the device structure. The composite passivation layer is deposited by ALD technology. Every 5 - 10 nm of Al 2 O 3 is deposited, and then nitrogen plasma with a power of 100 - 200W is introduced for treatment for 10 - 30 seconds. The nitrogen plasma reacts with the defects and impurities on the surface of the Al 2 O 3 layer, reducing the interfacial state density between layers to <1×10¹¹ cm⁻²⋅eV⁻¹, improving the electrical properties of the composite passivation layer, effectively reducing the surface leakage phenomenon, making the surface leakage current <0.5 nA, and at the same time increasing the breakdown field strength to >15 MV / cm. By repeatedly performing the ALD deposition and nitrogen plasma treatment steps, a composite passivation layer with excellent performance is formed, providing reliable protection for the device and improving the stability and reliability of the device. Further, when suppressing surface leakage, an improved sliding mode surface function is used to evaluate the leakage state, and the output of the improved sliding mode surface function is: In formula (1), represents the output of the sliding mode surface function, A represents the function stability parameter, represents the sliding mode surface function, represents the system state variable, u represents the continuity value of the sliding mode surface function, and b represents the state variable weight; where: In formula (2), i represents the number of leakage state types, n represents the total number of leakage state types, leakage state output function, represents the total set of leakage state output functions; the total set of functions is any one of the parameters of position, velocity, current, or voltage; represents the state variable adjustment amount; in the above formulas (1) and (2), the final output value of the sliding mode surface function It is used to quantitatively evaluate the surface leakage state of the electrostatic discharge module, and the numerical value directly reflects the degree of leakage or the suppression effect. The function stability parameter A is used to measure the stability of the output of the sliding mode surface function. The larger the value, the smaller the output fluctuation of the function when affected by external interference. It is the sliding mode surface function It is constructed by integrating different leakage state parameters and is used to describe the comprehensive state related to system leakage. System state variable Represents physical quantities related to leakage (such as current, voltage, and device position changes in a circuit), and directly reflects the system characteristics during leakage. The continuity value u of the sliding mode surface function reflects the continuous characteristic of the function mathematically and affects the smoothness of the control strategy. The higher the value, the more continuous the function output change. The state variable weight b is used to adjust The influence degree on the final output Reflects the importance of different state variables in leakage assessment. The total number n of leakage state types is the total number of all preset leakage state types and is the statistical upper limit of i. The i th leakage state output function Corresponds to the quantitative output of a specific leakage state (such as the current value function in the overcurrent state). The total set of leakage state output functions contains all preset leakage-related parameters (such as position, speed, current, voltage, etc.) and is the set. The state variable adjustment amount Is used to optimize the calculation. By adjusting the the sliding mode surface function can better fit the actual leakage state. The state variable adjustment amount is used to optimize the calculation. By adjusting the the sliding mode surface function can better fit the actual leakage state. In the specific implementation of the above formulas (1) and (2), first build a circuit experiment platform, set sensors in the electrostatic discharge module to collect leakage-related parameters (such as current and voltage) in real time. According to formula (2), count the total number n of leakage state types (such as setting 2 types of overcurrent and overvoltage, n = 2), and record the number i of actual leakage state types that occur (if only overcurrent is monitored, i = 1). Measure the i-th leakage state output function xi (such as the current value function during overcurrent), and combine the preset ci (determined through experimental debugging, such as c1 = 0.8) to calculate s(x)=c 1 x 1 +x 2 Finally, substitute it into formula (1), and combine A, b, u, f(p) (such as f(p) corresponding to the current current value) to calculate and obtain Secondly, repeat the experiment multiple times (such as 10 times) for the same leakage scenario, and record each time's Calculate the variance of the output value. The smaller the variance, the higher the stability of the function. Set A through fitting or empirical values (if the variance is small, take a larger value for A). Connect current sensors and voltage sensors in the circuit to monitor the current and voltage values during leakage in real time as the input of f(p). Analyze the continuity according to the mathematical expression of the sliding mode surface function (if the function has no mutation, take a high value for u), or adjust by observing the output smoothness through experiments. Use the trial-and-error method to adjust the value of b (for example, first set b = 0.5 and observe the response to f(p)), until it can accurately reflect the leakage state and determine the final value of b. Finally, according to the circuit design, preset the types of leakage states (such as overcurrent, overvoltage, and poor contact, n = 3); the number of actual leakage types that occur in the experiment (such as only overcurrent and overvoltage, i = 2); for each leakage state, obtain the corresponding parameters through sensors (such as x 1 is the current function during overcurrent, and x 2 is the voltage function during overvoltage); through multiple experiments for optimization, adjust to make s(x) more accurate. Further, the system state variable output function is: In formula (3), represents the state variable of the sliding mode variable structure control system; when calculating the system state variable output function, set the sliding mode surface function Use the Lyapunov function to determine the stability of the sliding mode surface function, and the output of the determination function is: Formula (4) is the expression of the Lyapunov function, and this function satisfies When the function enters the sliding mode surface, the system satisfies which indicates that the surface leakage suppression effect is better; in the above formulas (3) and (4), the working process of the sliding mode surface function is: First, the system collects the state variables of the sliding mode variable structure control system in real time (such as and related parameters) in formula (3), covering physical quantities such as current, voltage, and leakage parameters, as the basis for system state evaluation. Second, according to the collected state variables, substitute them into formula (2) to calculate s ( x ), integrate the leakage state type output function xi and the adjustment amount ci to form a comprehensive description of the leakage state. Finally, use the Lyapunov function formula (4) to determine the stability of the system; calculate and analyze its derivative indicating that the system converges to the sliding mode surface. When the system enters the sliding mode surface and satisfies At this time, the output of the sliding mode surface function is stable, and the surface leakage current suppression effect reaches an ideal state, triggering the corresponding control strategy to maintain this state. In a specific embodiment, the total number of leakage state types is preset in the software n , and the state variable adjustment amount is set c i (through preliminary experimental debugging, such as c 1 = 0.7,[[]] c 2 = 0.6), a sliding mode surface function is constructed At the same time, the calculation logic of the system state variable output function f ([[]] p ) is defined, and the mapping relationship between p and the actual physical quantities (such as the leakage current amplitude and the voltage change rate) is clarified. A data processing program is written to substitute the current and voltage data collected by the sensor into the formula to calculate f ([[]] p ) and s ([[]] x ). If an increase in the leakage current is detected, the program calculates the current system state variable output according to , and then combines it with s ([[]] x ) formula to integrate the leakage state. The calculation logic of the Lyapunov function and its derivative is realized. The software analyzes in real time to determine that the system converges to the sliding surface, triggering control optimization and adjustment. When the software determines that dtdV < 0, the microcontroller sends an instruction to the drive circuit. When it is detected that the surface leakage of the ESD protection device causes abnormal voltage, the drive circuit adjusts the field plate voltage to enhance the regulation of the electric field, so that the charge is evenly discharged and the local electric field concentration is suppressed. When the system satisfies , it indicates that it enters the sliding surface and the leakage current suppression effect is good. At this time, the software maintains the current control parameters (such as maintaining the field plate voltage value and the transistor conduction state), and continuously monitors S ([[]] x ) through closed-loop feedback. If an external interference (such as an instantaneous voltage fluctuation) is detected, s ([[]] x ) and V are recalculated, and the control is adjusted to ensure that the device is always in a stable state of leakage current suppression. During the electrostatic discharge process, the conduction degree of the discharge path is continuously optimized to enable the charge to be quickly discharged and maintain the high-voltage resistance performance of the protection device.

[0043] Furthermore, when discharging in the charge discharge area, the average values of the active power and the reactive power are as follows: In formula (5), It is the average value of the active power output of the ESD protection device; It is the average value of reactive power output by ESD protection device; is the voltage output by the first semiconductor layer (2) and the second semiconductor layer (3); is the sinusoidal current output by the first semiconductor layer (2) and the second semiconductor layer (3); in a specific embodiment, the output voltage of the first semiconductor layer 2 is obtained The second semiconductor layer 3 outputs a voltage and a corresponding sinusoidal current Multiply them separately and sum them, then multiply them by the coefficient -1.6 which is theoretically derived and experimentally verified to construct the active power mean. Accurately quantify the power value of energy converted into actual discharge energy during discharge. Multiply them separately and sum them, then multiply by the coefficient -1.6 to get the mean reactive power. Reflects the energy exchange characteristics of energy storage elements (such as semiconductor junction capacitors) in the device. When the bidirectional high-voltage ESD protection device is working, the first and second semiconductor layers bear the core functions of charge discharge and energy conversion. Under the impact of electrostatic pulses, the dynamic changes in the voltage and current output of the two layers affect the power output. The power function is defined to accurately describe the power output in the charge discharge region.

[0044] Through the above definition, the energy conversion and transmission in the charge discharge area, as well as the performance of ESD protection devices, can be analyzed more accurately.

[0045] The sliding surface function output of the sliding mode variable structure is: In formula (6), are sliding surface functions of the active power and reactive power of the electrostatic discharge module (101), respectively; is the preset value of active power; P is the real-time value of active power; is a preset reactive power value; Q is a real-time reactive power value; (I) is the leakage current output by the electrostatic discharge module (101); (V) is the voltage at the leakage point output by the electrostatic discharge module (101); (Q) is the charge accumulation of the electrostatic discharge module (101); When the electrostatic discharge module (101) suppresses surface leakage to the best state, the rate of change of S is zero. In a specific embodiment, in a bidirectional high-voltage ESD protection device, when the electrostatic discharge module 101 is working, its real-time active power value P, and its real-time reactive power value Q are monitored in real time, and compared with the preset value Comparison. The leakage current I, leakage point voltage V, charge accumulation Q, and adjustment parameter k output by the electrostatic discharge module are used to construct a sliding mode surface function (Equation 6). The sliding mode surface function couples the power deviation with the physical quantities related to leakage to form a control target. By adjusting the parameter k, the operating state of the electrostatic discharge module 101 is forced to converge to the sliding mode surface. When it is detected that the surface leakage causes an increase in power deviation, the sliding mode variable structure control is triggered to adjust the field plate voltage and the conduction degree of the discharge path (such as controlling the conduction state of the transistor) in the module, optimize the charge discharge path, and suppress the leakage. Through continuous iterative adjustment, the system enters the steady-state region defined by the sliding mode surface, achieving efficient suppression of surface leakage. When the electrostatic discharge module 101 reaches the optimal state of suppressing surface leakage, the active power deviation no longer changes. The system state described by the sliding mode surface function S has reached a steady state - physical quantities such as the leakage current I and charge accumulation Q reach a dynamic balance and no additional adjustment is required.

[0046] Embodiment 3 Preparation and Performance Verification of a Bidirectional High-Voltage-Resistant ESD Protection Device 1 During the experiment, the aluminum component ratio of the AlGaN layer is selected to be 25%, and the thickness of the Al 2 O 3 layer is 5 and the thickness of the SiNx layer is 500 nm. The doping element of the P-type doping region is magnesium Mg, and the doping concentration is distributed in a gradient from the heterojunction interface to the surface, with a range of 1×10¹⁸ cm −3 . The thicknesses of the first metal electrode and the second metal electrode are 200 nm. After manufacturing, the structural parameters of the prepared device are shown in Table 1. The aluminum component of the AlGaN layer is 25% - 40%, forming a strong polarization heterojunction with GaN, inducing a high-concentration 2DEG (carrier concentration 1.2×10¹³ cm⁻², mobility > 2000 cm² / V·s) at the interface, providing a low-resistance conduction path for electrostatic pulses and reducing the voltage drop. The intrinsic breakdown field strength of GaN reaches 3.3 MV / cm, which is 10 times that of silicon. Coupled with the high critical electric field of the AlGaN layer > 2 MV / cm, the device is endowed with natural high-voltage resistance ability. The P-type doping region (Mg gradient doping, concentration 1×10¹ 8 ~5×10¹ 9cm⁻³) forms a reverse PN junction with 2DEG, triggering avalanche breakdown under ESD high voltage, and directing energy to the ground through the ohmic contact electrode. The first P well 6, the second P well 9, and the third P well 12 are arranged in parallel, and a space charge region is formed by alternating P+ or N+ doping, which reduces the reverse electric field peak by 30% and increases the breakdown voltage to >180V. The field plate is connected to the source or gate, and by introducing an additional electric field, the depletion layer is widened, and the electric field concentration at the edge of the PN junction is suppressed, so that the electric field concentration coefficient is <1.5, avoiding premature breakdown. The reverse PN junction plus field plate further adjusts the electric field gradient, and cooperates with the P well structure to achieve an overall withstand voltage ≥60V. The ALD alternately deposits Al 2 O 3 (5~20nm) and SiNx (50~200nm), using high dielectric constant or low dielectric constant interface to adjust the surface electric field, and suppress the leakage current to <0.5nA. The cascade transistor in the passivation layer adjusts the electric field distribution in real time when overvoltage occurs, further reducing the risk of surface breakdown. The P-type doped region (concentration 2×10¹) is formed in the grooves on both sides of the heterojunction. 8 ~6×10¹ 9 cm⁻³), precisely controlling the PN junction depth and doping gradient. The Ti / Al / Ni / Au multilayer structure (thickness 610~800nm) forms a low-resistance ohmic contact (contact resistance <5×10 ⁻7 Ω・cm²). Conclusion: Ensure that ESD energy is discharged quickly, so bidirectional high withstand voltage can be achieved.

[0047] Among them, Al 2 O 3 The thickness of the SiNx layer is 20 nm, the thickness of the SiNx layer is 200 nm, the doping element of the P-type doping region is magnesium Mg, and the doping concentration is distributed in a gradient from the heterojunction interface to the surface, ranging from 5×10 19 cm −3 The thickness of the first metal electrode and the second metal electrode is 500 nm. The above method is still used to prepare the device structure parameters as shown in Table 2. Through the above method, it can be known that the parameters of the bidirectional high-voltage ESD protection device of the present application meet the voltage resistance standards.

[0048] First, the 4H-SiC substrate has a high breakdown field strength and can withstand high voltage, providing a good basic support for the device. The thickness of the GaN layer is 3.0 μm, and the thickness of the AlGaN layer is 25 nm. The high breakdown field strength characteristics of GaN and AlGaN are utilized to withstand high voltage. Secondly, the carrier concentration is 1.2×10 13cm⁻², with a relatively high mobility, provides a low-resistance conduction path for electrostatic pulses, helping to reduce the voltage drop. The doping element is Mg / Zn (atomic ratio 4:1), and the gradient concentration ranges from 1×10 18 cm⁻³ to 5×10¹ 9 cm⁻³. The gradient doping design helps to optimize the electric field distribution and increase the breakdown voltage. Finally, a Ti(20 nm) / Al(200 nm) / Ni(50 nm) / Au(100 nm) stacked structure is adopted, and a low-resistance ohmic contact is formed through annealing to ensure that the ESD energy can be quickly discharged.

[0049] Conclusion: The parameters of the bidirectional high-voltage-resistant ESD protection device meet the withstand voltage standard, which can effectively suppress surface leakage and electric field concentration, and optimize the electric field distribution inside the device.

[0050] Example 5 Preparation The present invention also adopts the following technical solution, a method for realizing a bidirectional high-voltage-resistant ESD protection device, including: Epitaxially grow a first semiconductor layer and a second semiconductor layer on a semiconductor substrate in sequence to form a heterojunction structure; Form grooves on both sides of the heterojunction through photolithography and etching processes, and form a P-type doping region in the grooves by ion implantation process, with a doping concentration of 1×10 18 ∼5×10 19 cm −3 ; Deposit a metal layer on the surface of the heterojunction, and form a first metal electrode and a second metal electrode through an annealing process, and achieve an ohmic contact with the 2DEG channel; Adopt the atomic layer deposition ALD process to alternately deposit a high dielectric constant material (AlO 3 ) and a low dielectric constant material SiNx on the surface of the device to form a composite passivation layer.

[0051] In the above embodiment, during the preparation of the epitaxial growth of the first semiconductor layer GaN and the second semiconductor layer AlGaN, a pulsed doping process is adopted, and silane SiH 4 gas is periodically injected during the growth process of the AlGaN layer. The injection pulse frequency is 10∼50Hz, and the pulse duty cycle is 20%∼80%, so that the AlGaN layer forms a longitudinal doping concentration gradient of 1×10 17 ∼5×10 18 cm −3 , and at the same time, control the growth temperature at 1050±5 ∘ °C, and the V / III ratio is 2000∼3000 to increase the carrier mobility of the 2DEG channel to >2000 cm 2 / V⋅s.

[0052] In the above embodiments, the ion implantation of the P-type doping region adopts a dual-element co-doping process, and the implantation dose is 5×10 14 ∼2×10 15 cm −2 , where the atomic ratio of magnesium Mg to zinc Zn is controlled at 3:1∼5:1, and rapid thermal annealing is carried out in a nitrogen atmosphere at 800∼950 ∘ °C for 30∼120 s. A gradient doping concentration distribution is formed from the surface of 5×10 19 cm −3 to the interface of 1×10 18 cm −3 , and the reverse breakdown voltage is increased to >150 V.

[0053] In the above embodiments, the annealing process of the metal electrode adopts a two-step method: The first stage: Anneal in a hydrogen atmosphere at 400∼500 °C for 5∼10 min to make the Ti / Al / Ni / Au stack react with the AlGaN surface to generate a TiN∼Al alloy transition layer; The second stage: Rapidly anneal in a nitrogen atmosphere at 750∼850 °C for 10∼30 s to reduce the specific contact resistance to <5×10 −7 Ω⋅cm2, while suppressing the thermal decomposition of the AlGaN layer.

[0054] In the ALD deposition process of the composite passivation layer, every time 5∼10 nm of Al 2 O 3 is deposited, 100∼200 W of nitrogen plasma is introduced for 10∼30 s to reduce the density of interface states between layers to <1×10 11 cm −2 ⋅eV −1 , the surface leakage current <0.5 nA, and the breakdown field strength >15 MV / cm.

[0055] Preparation process Step 1: Heterojunction epitaxial growth On a 4H-SiC substrate, grow a GaN layer by MOCVD process (1050 ∘ °C, V / III ratio = 2500); Pulsed doping of the AlGaN layer: Every time 5 nm of AlGaN is grown, inject SiH 4 gas (duty cycle 50%) at a pulse frequency of 30 Hz to form a longitudinal doping gradient (1×10 17 ∼3×10 18 cm −3 ).

[0056] Step 2: Formation of P-type doping region Lithography defines the groove region, and Mg / Zn is implanted by ion implantation process (dose 1×10 15 cm−2); Rapid thermal annealing: 900 ∘ C in nitrogen atmosphere, annealing time 60 s, to form a graded doped PN junction.

[0057] Step 3: Metal electrode preparation Deposit a Ti / Al / Ni / Au stack, two-step annealing: 1. Anneal in hydrogen at 450∘C for 8 min to form a TiN-Al transition layer; 2. Rapidly anneal in nitrogen at 800∘C for 20 s, specific contact resistance 4.2×10 −7 Ω⋅cm 2 .

[0058] Step 4: Composite passivation layer deposition Deposit Al 2 O 3 (10 nm) by ALD, and then introduce 150 W nitrogen plasma for 20 s; Repeat 3 times, total thickness 330 nm, interface state density 8×10 10 cm−2⋅eV−1.

[0059] The performance test data is shown in Table 3. Through experiments, during pulsed doping epitaxy, by dynamically modulating the SiH 4 pulse, the longitudinal doping gradient of AlGaN increases the 2DEG mobility to 2100 cm 2 / V⋅s. When double-element co-doping, the co-doping of Mg / Zn optimizes the carrier distribution, and the reverse breakdown voltage reaches 160 V. It has positive and beneficial technical effects.

[0060] Ultra-low contact resistance (<5×10⁻ 7 Ω⋅cm²) is achieved through staged annealing, and the thermal decomposition of AlGaN is avoided, ensuring the integrity of the heterojunction structure and providing a low-resistance path for the stable discharge of current during reverse breakdown. Nitrogen plasma treatment increases the breakdown field strength of the composite passivation layer (Al 2 O 3 / SiNx) to 16 MV / cm, effectively dispersing the peak electric field during reverse breakdown, suppressing surface leakage (<0.5 nA), and avoiding the reduction of reverse breakdown voltage caused by surface defects. The P-type region adopts a double-element (Mg / Zn) graded doping process, optimizing the electric field distribution of the PN junction and raising the theoretical upper limit of the reverse breakdown voltage. The two-way withstand voltage >600 V in the experimental conclusion directly proves the realization of the reverse breakdown voltage, covering both forward and reverse withstand voltage scenarios. Conclusion: In this embodiment, through an innovative process combination (pulse doping, dual-element gradient doping, two-step annealing, and plasma ALD), among which the breakdown voltage test includes reverse breakdown conditions, and the conclusion that the bidirectional voltage withstand > 600 V is precisely verified by testing the compliance of the reverse breakdown voltage. This indicates that from design, process to testing, all are carried out around the goal of bidirectional voltage withstand (including reverse breakdown voltage), achieving bidirectional voltage withstand > 600 V and ESD protection > 15 kV, and conducting electrical tests on the device.

[0061] Although the specific embodiments of the present invention have been described above, those skilled in the art should understand that these specific embodiments are merely illustrative. Without departing from the principles and essence of the present invention, those skilled in the art can make various omissions, substitutions, and changes to the details of the above methods and systems. For example, combining the above method steps, so as to perform substantially the same function in a substantially the same way to achieve substantially the same result, then it belongs to the scope of the present invention. Therefore, the scope of the present invention is only defined by the.

Claims

1. A bidirectional high-voltage ESD protection device, characterized in that: include: The semiconductor substrate (1) serves as the basic supporting structure of the entire device and provides a platform for the construction of other components; An electrostatic discharge module (101) is provided on the semiconductor substrate (1); A heterojunction structure formed on the semiconductor substrate (1), wherein a protection module is arranged on the heterojunction structure, wherein the heterojunction structure comprises a first semiconductor layer (2) and a second semiconductor layer (3), and a two-dimensional electron gas (2DEG) conductive channel is formed at the interface of the heterojunction structure; and the first semiconductor layer (2) and the second semiconductor layer (3) are both multi-level stacked structures; wherein the first semiconductor layer (2) is gallium nitride (GaN), and the second semiconductor layer (3) is aluminum gallium nitride (AlGaN), and the polarization effect at the heterojunction interface forms a two-dimensional electron gas channel, and the aluminum component of the AlGaN layer accounts for 25% to 40%; A first metal electrode (131) and a second metal electrode (132) symmetrically distributed on both sides of the heterojunction structure, respectively forming ohmic contact with the 2DEG channel; The P-type doped regions located on both sides of the heterojunction structure are formed by ion implantation or epitaxial growth and form a reverse PN junction with the 2DEG channel; the reverse PN junction is provided with a first P well (6) formed by a first P+ heavily doped region (4) and a first N+ heavily doped region (5), a second P well (9) formed by a second N+ heavily doped region (7) and a second P+ heavily doped region (8), and a third P well (12) formed by a third P+ heavily doped region (10) and a third N+ heavily doped region (11); wherein the first P well (6), the second P well (9) and the third P well (12) are arranged in parallel; the P-type doped region is provided with a reinforced metal field plate, which is connected to a source or a gate; and the reverse PN junction is also provided with an additional field plate; A composite passivation layer covering the heterojunction structure and the doped region, wherein the composite passivation layer comprises at least two layers of materials with different dielectric constants, and is used to suppress surface leakage and electric field concentration; the trigger voltage range is ±20V to ±20V, and the withstand voltage capability is ≥60V.

2. The bidirectional high voltage ESD protection device according to claim 1, characterized in that: The electrostatic discharge module (101) comprises a first relay switch circuit (19) and a second relay switch circuit (20) connected in parallel, the first relay switch circuit (19) being connected in series with a first transistor circuit (21) and a second transistor circuit (22) connected in series with each other, the second relay switch circuit (20) being connected in series with a third transistor circuit (23) and a fourth transistor circuit (24) connected in series with each other, the second transistor circuit (22) and the third transistor circuit (23) being connected in parallel, the output end of the third transistor circuit (23) being connected in series with a first current conversion circuit (25), and the output end of the fourth transistor circuit (24) being connected in series with a second current conversion circuit (26).

3. The bidirectional high voltage ESD protection device according to claim 1, characterized in that: The protection module is a voltage sensing circuit based on a high voltage transformation ratio of 1:10; the output end of the voltage sensing circuit is connected to the input end of the diode protection circuit, the output end of the diode protection circuit is connected to the input end of the discharge circuit, and the output end of the discharge circuit is connected to the input end of the output circuit.

4. The bidirectional high voltage ESD protection device according to claim 1, characterized in that: The composite passivation layer is composed of a high dielectric constant material Al2O3 and a low dielectric constant material SiNx deposited alternately, wherein the thickness of the Al2O3 layer is 5 to 20 nm, and the thickness of the SiNx layer is 50 to 200 nm; and wherein the composite passivation layer comprises a first passivation layer (14) and a second passivation layer (18) connected via a second transistor (16) arranged in parallel; a first transistor (15) is arranged in the first passivation layer (14), and a third transistor (17) is arranged in the second passivation layer (18); and the first transistor (15) is connected to the third transistor (17) via the second transistor (16) in a cascade structure.

5. The bidirectional high voltage ESD protection device according to claim 1, characterized in that: The doping element of the P-type doping region is magnesium Mg, and the doping concentration is distributed in a gradient from the heterojunction interface to the surface, ranging from 1×10 18 cm −3 Up to 5×10 19 cm −3 .

6. The bidirectional high voltage ESD protection device according to claim 1, characterized in that: The first metal electrode (131) and the second metal electrode (132) constitute a charge release region; the first metal electrode (131) and the second metal electrode (132) have a thickness of 610 to 800 nm and comprise a titanium Ti or aluminum Al or nickel Ni or gold Au stacked structure.

7. A method for realizing a bidirectional high-voltage ESD protection device, characterized in that: A bidirectional high-voltage ESD protection device according to any one of claims 1 to 6, comprising: Step 1: preparing a semiconductor substrate (1), and sequentially epitaxially growing a first semiconductor layer (2) and a second semiconductor layer (3) on the semiconductor substrate to form a heterojunction structure; the first semiconductor layer (2) is gallium nitride GaN, the second semiconductor layer (3) is aluminum gallium nitride AlGaN, and the aluminum component of the AlGaN layer accounts for 25% to 40%, and a two-dimensional electron gas 2DEG conductive channel is formed at the interface of the heterojunction structure; Step 2: arranging a protection module on the heterojunction structure, and depositing a metal layer on the surface of the heterojunction, forming a first metal electrode (131) and a second metal electrode (132) by an annealing process, and achieving ohmic contact with the 2DEG channel; forming a P-type doped region on both sides of the heterojunction structure, which is formed by ion implantation or epitaxial growth, and forms a reverse PN junction with the 2DEG channel; Step 3: a first P+ heavily doped region (4), a first N+ heavily doped region (5), a second N+ heavily doped region (7), a second P+ heavily doped region (8), a third P+ heavily doped region (10) and a third N+ heavily doped region (11) are arranged in the reverse PN junction to form a first P well (6), a second P well (9) and a third P well (12), wherein the first P well (6), the second P well (9) and the third P well (12) are arranged in parallel; an enhanced metal field plate is arranged in the P-type doped region, the field plate is connected to the source or the gate; and an additional field plate is arranged on the reverse PN junction; Step 4: Covering the heterojunction structure and the doped region with a composite passivation layer, wherein the composite passivation layer comprises at least two layers of materials with different dielectric constants, for suppressing surface leakage and electric field concentration; Grooves are formed on both sides of the heterojunction by photolithography and etching processes, and P-type doped regions are formed in the grooves by ion implantation, with a doping concentration of 2×10 18 ∼6×10 19 cm −3 ; The atomic layer deposition (ALD) process is used to alternately deposit high dielectric constant material Al2O3 and low dielectric constant material SiNx on the device surface to form a composite passivation layer.

8. The method for realizing a bidirectional high-voltage ESD protection device according to claim 7, characterized in that: The epitaxially grown first semiconductor layer (2) is GaN, and the second semiconductor layer (3) is AlGaN. During the application process, a pulsed doping process is used to periodically inject silane SiH4 gas during the growth of the AlGaN layer, with an injection pulse frequency of 10 to 50 Hz and a pulse duty cycle of 20% to 80%, so that the AlGaN layer forms a longitudinal doping concentration gradient of 1×10 17 ∼5×10 18 cm −3 , while controlling the growth temperature at 1050±5 ∘ C, V / III ratio is 2000∼3000 to increase the carrier mobility of 2DEG channel to >2000cm 2 / V⋅s.

9. The method for realizing a bidirectional high voltage ESD protection device according to claim 7, characterized in that: The ion implantation of the P-type doping region adopts a dual-element co-doping process, and the implantation dose is 5×10 14 ∼2×10 15 cm −2 The atomic ratio of magnesium Mg to zinc Zn is controlled at 3:1 to 5:1, and the ∘ C in a nitrogen atmosphere for rapid thermal annealing for 30 to 120 s. 19 cm −3 To interface 1×10 18 cm −3 Forming a gradient doping concentration distribution, the reverse breakdown voltage is increased to >180V; The annealing process of the metal electrode adopts a two-step method: The first stage: annealing in a hydrogen atmosphere at 400-500°C for 5-10 min to allow the Ti / Al / Ni / Au stack to react with the AlGaN surface to form a TiN-Al alloy transition layer; The second stage: rapid annealing in a nitrogen atmosphere at 750-850°C for 10-30s to reduce the specific contact resistance to <5×10 −7 Ω⋅cm2, while suppressing the thermal decomposition of the AlGaN layer; During the ALD deposition of the composite passivation layer, after each 5-10 nm Al2O3 deposition, a 100-200 W nitrogen plasma treatment was conducted for 10-30 s to reduce the interlayer interface state density to <1×10 11 cm −2 ⋅eV −1 , surface leakage current <0.5nA, breakdown field strength >15MV / cm.

10. The method for realizing a bidirectional high voltage ESD protection device according to claim 7, characterized in that: When suppressing surface leakage, an improved sliding surface function is used to evaluate the leakage state, wherein the output of the improved sliding surface function is: In formula (1), represents the output of the sliding surface function, A represents the function stability parameter, represents the sliding surface function, represents the system state variable, u represents the continuity value of the sliding surface function, and b represents the state variable weight; where: In formula (2), i represents the number of leakage state types, n represents the total number of leakage state types, Leakage status output function, Indicates a total set of output functions of leakage state; the total set of functions is any one parameter of position, speed, current or voltage; Represents the state variable adjustment; the system state variable output function is: In formula (3), Represents the state variables of the sliding mode variable structure control system; when calculating the output function of the system state variables, the sliding surface function is assumed The Lyapunov function is used to determine the stability of the sliding surface function. The output of the determination function is: Formula (4) is the expression of the Lyapunov function, which satisfies The function enters the sliding surface, and the system satisfies It indicates that the surface leakage suppression effect is better; When the charge discharge region is discharged, the average values ​​of active power and reactive power are as follows: In formula (5), It is the average value of the active power output of the ESD protection device; It is the average value of reactive power output by ESD protection device; is the voltage output by the first semiconductor layer (2) and the second semiconductor layer (3); is the sinusoidal current output by the first semiconductor layer (2) and the second semiconductor layer (3); the sliding mode surface function output of the sliding mode variable structure is: In formula (6), are sliding surface functions of the active power and reactive power of the electrostatic discharge module (101), respectively; is the preset value of active power; P is the real-time value of active power; is a preset reactive power value; Q is a real-time reactive power value; (I) is the leakage current output by the electrostatic discharge module (101); (V) is the voltage at the leakage point output by the electrostatic discharge module (101); (Q) is the charge accumulation of the electrostatic discharge module (101); When the electrostatic discharge module (101) suppresses surface leakage to an optimal state, the rate of change of S is zero.

Images