Radio frequency semiconductor device, electronic device, and method for manufacturing radio frequency semiconductor device

Abstract

Embodiments of the present application provide a radio frequency semiconductor device, an electronic device and a preparation method of the radio frequency semiconductor device. It relates to the technical field of semiconductors. A radio frequency semiconductor device with low radio frequency loss is provided. The radio frequency semiconductor device comprises a substrate, a nucleation layer, a channel layer and a barrier layer stacked in sequence on the substrate, wherein the nucleation layer, the channel layer and the barrier layer all contain elements in group IIIA; the nucleation layer further contains P-type impurities, the P-type impurities contain at least one element in group II and / or group IV A, and the nucleation layer comprises at least two layers stacked. By introducing at least two layers of the nucleation layer containing P-type impurities, the formation of a parasitic channel in the substrate containing silicon is inhibited, and the radio frequency loss of the radio frequency semiconductor device is reduced.

Description

Technical Field

[0001] This application relates to the field of semiconductor technology, and in particular to a radio frequency semiconductor device, an electronic device having the radio frequency semiconductor device, and a method for fabricating the radio frequency semiconductor device. Background Technology

[0002] With the development of fifth-generation wireless communication technologies (5G), the field of radio frequency communication has put forward requirements for semiconductor devices with higher frequencies, higher voltages, higher output power, and higher efficiency.

[0003] Semiconductor devices based on compound semiconductor materials, such as gallium nitride (GaN)-based high electron mobility transistors (HEMTs), are increasingly being used in high-power radio frequency devices, high-voltage switching devices, and other fields due to their high thermal conductivity, high breakdown field strength, and high saturation electron mobility. For example, they have wide applications in radar, wireless communication, navigation, satellite communication, and electronic countermeasures equipment systems.

[0004] Figure 1 The diagram shown is a structural diagram of a radio frequency semiconductor device. See Figure 1 As shown, the radio frequency semiconductor device includes a silicon substrate, a GaN layer formed on the silicon substrate, an aluminum gallium nitride (AlGaN) layer formed on the GaN layer, and a source, drain, and gate formed on the AlGaN layer. Other functional film structures may also be stacked between the silicon substrate and the GaN layer, between the AlGaN layer and the electrodes (including the source, drain, and gate), or between the GaN layer and the AlGaN layer.

[0005] For radio frequency (RF) applications, a key requirement for RF semiconductor devices is to reduce conduction losses during RF operation, i.e., RF losses. For example, when used in RF power amplifiers (PAs), low RF losses can improve the output power, power gain, and efficiency of the power amplifier. However, in... Figure 1 In the structure shown, parasitic conductive channels (also called parasitic channels) are easily formed on the surface of the silicon substrate. These parasitic conductive channels will generate radio frequency losses and reduce the operating performance of the radio frequency semiconductor.

[0006] Therefore, how to provide a radio frequency semiconductor device with low radio frequency loss is a technical problem that urgently needs to be solved. Summary of the Invention

[0007] This application provides a radio frequency (RF) semiconductor device, an electronic device having the RF semiconductor device, and a method for fabricating the RF semiconductor device. By providing a novel RF semiconductor device, low RF loss is achieved.

[0008] To achieve the above objectives, the embodiments of this application adopt the following technical solutions:

[0009] On one hand, this application provides a radio frequency semiconductor device. For example, the radio frequency semiconductor device may be a gallium nitride (GaN) based high electron mobility transistor (HEMT).

[0010] The radio frequency semiconductor device includes: a substrate, a nucleation layer, a channel layer, and a barrier layer; wherein the substrate comprises silicon, and the nucleation layer, the channel layer, and the barrier layer are sequentially stacked on the substrate; the nucleation layer, the channel layer, and the barrier layer all contain group IIIA nitrides; the nucleation layer includes a first nucleation layer and a second nucleation layer stacked in a first direction, the first nucleation layer being formed on the substrate, and the second nucleation layer being formed on the first nucleation layer; both the first nucleation layer and the second nucleation layer contain p-type impurities, and the p-type impurities contain at least one element from group II and / or group IVA; the concentration of p-type impurities in the first nucleation layer is greater than the concentration of p-type impurities in the second nucleation layer; and / or, the p-type impurities in the first nucleation layer and the p-type impurities in the second nucleation layer are different. That is, in some embodiments, the concentration of P-type impurities in the first nucleation layer is greater than the concentration of P-type impurities in the second nucleation layer; in other embodiments, the P-type impurities in the first nucleation layer and the second nucleation layer are different. For example, when the first nucleation layer and the second nucleation layer contain only one type of P-type impurity, the P-type impurities involved are different; for another example, when the first nucleation layer and the second nucleation layer each contain at least two types of P-type impurities, at least some of the P-type impurities are different; in still other embodiments, not only is the concentration of P-type impurities in the first nucleation layer greater than the concentration of P-type impurities in the second nucleation layer, but the P-type impurities in the first nucleation layer and the second nucleation layer are also different.

[0011] In the radio frequency semiconductor device disclosed in this application, the nucleation layer includes not only a first nucleation layer and a second nucleation layer stacked in a first direction, but also both the first nucleation layer and the second nucleation layer contain P-type impurities. Thus, the band structure between the acceptor nucleation layer and the substrate suppresses the formation of N-type parasitic channels in the inversion layer on the surface of the substrate, reducing the conductivity of the substrate and thereby reducing the radio frequency loss of the radio frequency semiconductor device.

[0012] Furthermore, in this application, the concentration of P-type impurities in the first nucleation layer is greater than the concentration of P-type impurities in the second nucleation layer. That is, the first nucleation layer, containing a higher concentration of P-type impurities, is positioned closer to the substrate than the second nucleation layer, which contains a lower concentration of P-type impurities. This allows the first nucleation layer to suppress the formation of N-type parasitic channels in the inversion layer on the substrate surface, and also prevents the formation of parasitic channels within the nucleation layer itself due to high concentrations of P-type impurities in the second nucleation layer. Consequently, the RF loss of the RF semiconductor device can be further reduced.

[0013] In one feasible manner, along a first direction, the thickness of the first nucleating layer is less than the thickness of the second nucleating layer.

[0014] In one feasible approach, the thickness d1 of the first nucleation layer is: 1nm≤d1≤100nm; and the thickness d2 of the second nucleation layer is: 100nm≤d2≤250nm.

[0015] Of course, in some other structures, the thickness of the first nucleating layer can be greater than or equal to the thickness of the second nucleating layer along the stacking direction.

[0016] In one feasible approach, the dislocation density of the second nucleation layer is less than that of the first nucleation layer.

[0017] Since the second nucleation layer is located further away from the substrate than the first nucleation layer, and the dislocation density of the second nucleation layer is smaller than that of the first nucleation layer, the lattice arrangement of the channel layer can be made more orderly when preparing subsequent film structures, such as channel layers, thereby improving the electron mobility of the channel layer.

[0018] In one feasible manner, both the P-type impurities in the first nucleation layer and the P-type impurities in the second nucleation layer contain at least two elements.

[0019] In one feasible approach, the half-width at half-maximum (FWHM) of the rocking curve of the nucleation layer is less than or equal to 800 arcseconds.

[0020] Compared to existing nucleation layers with a half-width at half-maximum (WHM) of more than 1000 arcseconds, the radio frequency semiconductor device presented in this application has a lower dislocation density, thereby improving the device's performance and reliability.

[0021] In one feasible implementation, the concentration ρ of the P-type impurities in both the first and second nucleation layers is 1 × 10⁻⁶. 15 cm -3 ≤ρ≤1×10 20 cm -3 .

[0022] In one feasible manner, the concentration of P-type impurities in the first nucleation layer is ρ1; the concentration of P-type impurities in the second nucleation layer is ρ2; wherein ρ1 / ρ2 > 10.

[0023] In one feasible approach, the concentration ρ1 of the P-type impurities in the first nucleation layer is: 1 × 10⁻⁶. 17 cm -3 ≤ρ≤1×10 20 cm -3 The concentration ρ2 of the P-type impurities in the second nucleation layer is: 1 × 10⁻⁶ 15 cm -3 ≤ρ≤1×10 18 cm -3 .

[0024] In one feasible manner, P-type impurities include at least one of beryllium, magnesium, calcium, zinc, carbon, and silicon.

[0025] In one possible implementation, the P-type impurities in the first nucleation layer include magnesium; the P-type impurities in the second nucleation layer include at least two of magnesium, carbon, and silicon.

[0026] In one possible implementation, the Group IIIA nitride in the first nucleation layer is the same as the Group IIIA nitride in the second nucleation layer.

[0027] From a process perspective, it is easier to prepare the first and second nucleation layers using the same Group IIIA nitrides.

[0028] In one possible implementation, both the first nucleation layer and the second nucleation layer comprise aluminum nitride (AlN).

[0029] P-type impurities are doped into a Group IIIA nitride nucleation layer containing aluminum nitride (AlN). In this layer, the lattice arrangement is not very regular due to lattice mismatch with the substrate, resulting in a large number of dislocations. The purpose of the Group IIIA nitride nucleation layer is to reduce the dislocation density of the channel layer, making the lattice arrangement of the channel layer grown thereon more regular and improving its performance. In this embodiment, doping P-type impurities into the less regularized Group IIIA nitride nucleation layer does not severely disrupt the lattice arrangement, thus having minimal impact on its performance.

[0030] In one possible implementation, the channel layer comprises gallium nitride (GaN) material.

[0031] This can be understood as follows: the group IIIA nitride nucleation layer containing aluminum nitride (AlN) material can ensure a more orderly lattice arrangement in the group IIIA nitride channel layer containing gallium nitride (GaN) material, thereby improving the mobility of the two-dimensional electron gas (2DEG) in the group IIIA nitride channel layer.

[0032] In one possible implementation, the barrier layer comprises aluminum gallium nitride (AlGaN) material.

[0033] For example, a group IIIA nitride barrier layer is placed on top of a group IIIA nitride channel layer. The group IIIA nitride barrier layer is used to cooperate with the group IIIA nitride channel layer to generate 2DEG through polarization in the region where the group IIIA nitride channel layer and the group IIIA nitride barrier layer meet, thereby providing a channel for conducting current.

[0034] In one possible implementation, the radio frequency semiconductor device further includes: a diffusion barrier layer; the diffusion barrier layer is formed between the substrate and the first nucleation layer along a first direction.

[0035] By setting a diffusion barrier layer between the substrate and the first nucleation layer, the diffusion barrier layer can be used to suppress the diffusion of Group IIIA elements in the nucleation layer or other film structures located on the nucleation layer into the substrate, thereby suppressing the formation of P-type parasitic channels on the surface of the substrate and further reducing the conductivity of the substrate.

[0036] On the other hand, this application also provides a method for fabricating a radio frequency semiconductor device, the method comprising:

[0037] A first nucleation layer is formed on a silicon-containing substrate, the first nucleation layer comprising a group IIIA nitride doped with a p-type impurity;

[0038] A second nucleation layer is formed on the first nucleation layer, the second nucleation layer comprising a Group IIIA nitride doped with a P-type impurity;

[0039] P-type impurities contain at least one element from Group II and / or Group IVA.

[0040] The concentration of P-type impurities in the first nucleation layer is greater than that in the second nucleation layer; and / or, the concentration of P-type impurities in the first nucleation layer is different from that in the second nucleation layer.

[0041] In the method for fabricating a radio frequency semiconductor device provided in this application, since the group IIIA nitrides forming the first nucleation layer and the second nucleation layer are both doped with P-type impurities, the first nucleation layer and the second nucleation layer containing P-type impurities can suppress the formation of N-type parasitic channels of the inversion layer on the surface of the substrate, thereby reducing the conductivity of the substrate and reducing the radio frequency loss of the fabricated radio frequency semiconductor device.

[0042] In addition, in the fabrication method of this application, the concentration of P-type impurities in the first nucleation layer is greater than the concentration of P-type impurities in the second nucleation layer. That is, the first nucleation layer containing a higher concentration of P-type impurities is disposed closer to the substrate than the second nucleation layer containing a lower concentration of P-type impurities. In this way, the first nucleation layer can suppress the formation of N-type parasitic channels of the inversion layer on the surface of the substrate, and it will not cause parasitic channels to be generated in the nucleation layer itself due to the high concentration of P-type impurities in the second nucleation layer, thereby further reducing the RF loss of the RF semiconductor device.

[0043] In one feasible manner, during the formation of the first nucleation layer and the formation of the second nucleation layer, the growth temperature for forming the first nucleation layer is lower than the growth temperature for forming the second nucleation layer.

[0044] When the temperature at which the first nucleation layer grows is lower than the temperature at which the second nucleation layer grows, it helps to avoid excessive tensile stress in the first nucleation layer, thereby reducing the risk of crack formation.

[0045] In one feasible manner, the precursor for preparing the first nucleation layer and / or the second nucleation layer includes at least one of diethylberyllium, magnesium pyrocene, dimethylzinc, calcium diisopropyldithiocarbamate, methane, ethylene, silane, or disilane.

[0046] In one feasible manner, before forming the first nucleation layer on the silicon-containing substrate, the fabrication method further includes forming a diffusion barrier layer on the silicon-containing substrate to form the first nucleation layer on the diffusion barrier layer.

[0047] In the selectable process flow, a diffusion barrier layer can be formed on the substrate first, and then a first nucleation layer can be formed on the diffusion barrier layer. That is, the diffusion barrier layer can suppress the diffusion of group IIIA elements in the nucleation layer into the substrate, so as to form a P-type parasitic channel with conductivity in the substrate.

[0048] In one feasible approach, the same Group IIIA nitride is used to prepare both the first and second nucleation layers during their formation. For example, aluminum nitride (AlN) can be used to prepare both the first and second nucleation layers.

[0049] This ensures that the main components of the first nucleation layer and the second nucleation layer are the same.

[0050] In another aspect, this application also provides a radio frequency semiconductor device, which includes a substrate, a diffusion barrier layer, a nucleation layer, a channel layer, and a barrier layer; wherein, the substrate comprises silicon, and the diffusion barrier layer, the nucleation layer, the channel layer, and the barrier layer are stacked sequentially on the substrate; the nucleation layer, the channel layer, and the barrier layer all contain group IIIA nitrides; the nucleation layer also contains p-type impurities, and the p-type impurities contain at least one element from group II and / or group IVA.

[0051] In the radio frequency semiconductor device provided in this application, a diffusion barrier layer is formed between the substrate and the nucleation layer. This diffusion barrier layer can prevent group IIIA elements in the nucleation layer or other film structures located on the nucleation layer from diffusing into the substrate, thereby suppressing the formation of P-type parasitic channels on the surface of the substrate and reducing the conductivity of the substrate.

[0052] In addition, since the nucleation layer in this radio frequency semiconductor device also contains P-type impurities, the band structure between the acceptor nucleation layer and the substrate will suppress the formation of N-type parasitic channels of the inversion layer on the surface of the substrate, reduce the conductivity of the substrate, and thus reduce the radio frequency loss of the radio frequency semiconductor device.

[0053] In one feasible approach, the concentration ρ of P-type impurities in the nucleation layer is always 1 × 10⁻⁶. 15 cm -3 ≤ρ≤1×10 20 cm -3 .

[0054] In one feasible approach, the thickness d of the nucleation layer is: 1nm ≤ d ≤ 300nm.

[0055] For example, the thickness d of the nucleation layer can be equal to 150nm.

[0056] In one feasible manner, the P-type impurities in the nucleation layer include at least one of magnesium, carbon, and silicon.

[0057] In one possible implementation, the nucleation layer includes a first nucleation layer and a second nucleation layer stacked along a first direction, the first nucleation layer being formed on a diffusion barrier layer, the second nucleation layer being formed on the first nucleation layer, and both the first and second nucleation layers containing P-type impurities.

[0058] That is, the radio frequency semiconductor device may include at least two nucleation layers.

[0059] In one achievable manner, the concentration of P-type impurities in the first nucleation layer is greater than the concentration of P-type impurities in the second nucleation layer; and / or, the P-type impurities in the first nucleation layer and the P-type impurities in the second nucleation layer are different.

[0060] The concentration of P-type impurities in the first nucleation layer is greater than that in the second nucleation layer. That is, the first nucleation layer, containing a higher concentration of P-type impurities, is positioned closer to the substrate than the second nucleation layer, which contains a lower concentration of P-type impurities. This allows the first nucleation layer to suppress the formation of N-type parasitic channels in the inversion layer on the substrate surface, and also prevents the formation of high-conductivity parasitic channels within the nucleation layer itself due to the high concentration of P-type impurities in the second nucleation layer, thereby further reducing the radio frequency loss of the RF semiconductor device.

[0061] In one feasible approach, the concentration ρ of the P-type impurity in the first nucleation layer is: 1 × 10⁻⁶ 17 cm -3 ≤ρ≤1×10 20 cm -3 The concentration ρ of the P-type impurity in the second nucleation layer is: 1 × 10⁻⁶ 15 cm -3 ≤ρ≤1×10 18 cm -3 .

[0062] In one possible implementation, the P-type impurities in the first nucleation layer include magnesium; the P-type impurities in the second nucleation layer include at least two of magnesium, carbon, and silicon.

[0063] In one possible implementation, the Group IIIA nitride in the first nucleation layer is the same as the Group IIIA nitride in the second nucleation layer.

[0064] In another aspect, this application also provides a method for fabricating a radio frequency semiconductor device, the method comprising:

[0065] A diffusion barrier layer is formed on a substrate containing silicon;

[0066] A nucleation layer is formed on the diffusion barrier layer; wherein the nucleation layer further contains a P-type impurity, and the P-type impurity contains at least one element from group II and / or group IVA.

[0067] In the fabrication method described in this application, a diffusion barrier layer can prevent the diffusion of Group IIIA elements from the nucleation layer into the substrate, thereby suppressing the formation of P-type parasitic channels on the substrate surface and reducing the conductivity of the substrate. Furthermore, the nucleation layer also contains P-type impurities. The band structure between the acceptor nucleation layer and the substrate suppresses the formation of N-type parasitic channels in the inversion layer on the substrate surface, further reducing the substrate conductivity and thus lowering the radio frequency loss of the RF semiconductor device.

[0068] In one possible implementation, forming the nucleation layer includes: forming a first nucleation layer on a silicon-containing substrate, and then forming a second nucleation layer on the first nucleation layer, wherein the concentration of P-type impurities in the first nucleation layer is greater than the concentration of P-type impurities in the second nucleation layer; and / or, the P-type impurities in the first nucleation layer and the P-type impurities in the second nucleation layer are different.

[0069] In one feasible manner, the precursor for preparing the first nucleation layer and / or the second nucleation layer comprises at least one of diethylberyllium, magnesium pyrocene, dimethylzinc, calcium diisopropyldithiocarbamate, methane, ethylene, silane, or disilane.

[0070] In another aspect, this application also provides an electronic device, including a circuit board and a radio frequency semiconductor device in any of the above implementations or a radio frequency semiconductor device made in any of the above implementations, wherein the circuit board is electrically connected to the radio frequency semiconductor device.

[0071] The electronic device provided in this application includes the radio frequency semiconductor device described above. Therefore, the electronic device provided in this application and the radio frequency semiconductor device of the above technical solution can solve the same technical problem and achieve the same expected effect. Attached Figure Description

[0072] Figure 1 This is a schematic diagram of the structure of a radio frequency semiconductor device in the prior art;

[0073] Figure 2 This is a partial structural diagram of a base station;

[0074] Figure 3 An exploded view of part of the mobile phone's structure;

[0075] Figure 4 This is a schematic diagram of a part of the structure of a base station or mobile phone or other electronic device.

[0076] Figure 5 This is a schematic diagram of the structure of a radio frequency semiconductor device;

[0077] Figure 6 It shows Figure 5 A schematic diagram illustrating the principle of forming P-type parasitic channels in the structure shown.

[0078] Figure 7 This is a schematic diagram of the structure of a radio frequency semiconductor device provided in an embodiment of this application;

[0079] Figure 8 A schematic diagram illustrating the structure of an electrode in a radio frequency semiconductor device provided in this application embodiment;

[0080] Figure 9This is a schematic diagram of the structure of a radio frequency semiconductor device provided in an embodiment of this application;

[0081] Figure 10a The energy band diagrams of AlN and Si in a radio frequency semiconductor device provided in this application embodiment;

[0082] Figure 10b The present application provides an energy band diagram of AlN and Si, which are n-type conductivity types, in a radio frequency semiconductor device.

[0083] Figure 10c An energy band diagram of a heterojunction formed by AlN and Si of n-type conductivity in a radio frequency semiconductor device provided in an embodiment of this application;

[0084] Figure 11 This is a schematic diagram of the structure of a radio frequency semiconductor device provided in an embodiment of this application;

[0085] Figure 12a The present application provides an energy band diagram of AlN and Si, P-type conductivity types, in an embodiment of a radio frequency semiconductor device.

[0086] Figure 12b An energy band diagram of a heterojunction formed by AlN and Si of P-type conductivity in a radio frequency semiconductor device provided in this application embodiment;

[0087] Figure 13 This is a schematic diagram of the structure of a radio frequency semiconductor device provided in an embodiment of this application;

[0088] Figure 14 This is a schematic diagram of the structure of a radio frequency semiconductor device provided in an embodiment of this application;

[0089] Figure 15 This is a schematic diagram of the structure of a radio frequency semiconductor device provided in an embodiment of this application;

[0090] Figure 16 This is a schematic diagram of the structure of a radio frequency semiconductor device provided in an embodiment of this application;

[0091] Figure 17 This is a schematic diagram of the structure of a radio frequency semiconductor device provided in an embodiment of this application;

[0092] Figure 18 This is a schematic diagram of the structure of a radio frequency semiconductor device provided in an embodiment of this application;

[0093] Figure 19 This is a schematic diagram of the structure of a radio frequency semiconductor device provided in an embodiment of this application;

[0094] Figure 20This is a schematic diagram of the structure of a radio frequency semiconductor device provided in an embodiment of this application;

[0095] Figure 21 This is a schematic diagram of the structure of a radio frequency semiconductor device provided in an embodiment of this application;

[0096] Figure 22 A flowchart illustrating a method for forming a radio frequency semiconductor device according to an embodiment of this application;

[0097] Figures 23a to 23e A schematic diagram of the structure after each step in a method for manufacturing a radio frequency semiconductor device provided in this application embodiment is completed;

[0098] Figure 24 A flowchart illustrating a method for forming a radio frequency semiconductor device according to an embodiment of this application;

[0099] Figures 25a to 25f This is a schematic diagram showing the structure after each step of a method for manufacturing a radio frequency semiconductor device provided in this application embodiment is completed.

[0100] Figure label:

[0101] 11-Middle frame; 110-Border edge; 111-Support plate;

[0102] 12-Rear shell;

[0103] 13-Display screen;

[0104] 100 - Circuit board;

[0105] 101 - Radio frequency semiconductor device; 102 - First electrical connection structure; 103 - Packaging substrate;

[0106] 200 - Second electrical connection structure;

[0107] 1-Substrate;

[0108] 2-Diffusion barrier layer;

[0109] 31-Nucleation layer;

[0110] 311 - First nucleation layer; 312 - Second nucleation layer; 313 - Third nucleation layer;

[0111] 32-channel layer;

[0112] 33 - Barrier layer;

[0113] 34 - Stress buffer layer;

[0114] 35 - Isolation layer;

[0115] 36-hat layer

[0116] 41-Source;

[0117] 42-Gate;

[0118] 43-Drain electrode. Detailed Implementation

[0119] Before introducing the embodiments involved in this application, some technical terms related to the embodiments of this application will be introduced as follows:

[0120] Dislocation density: A large number of dislocations exist in ordinary crystals, and the quantity of these dislocations is expressed by dislocation density. Dislocation density can be defined as the total length of dislocation lines contained in a unit volume of crystal.

[0121] Half-width at half-maximum (WHM) of a rocking curve: This describes the angular divergence of a specific crystal plane in a sample. The larger the angular divergence, the larger the WHM of the rocking curve and the higher the dislocation density. Conversely, the smaller the angular divergence, the smaller the WHM of the rocking curve and the lower the dislocation density.

[0122] Precursor: This can be understood as the original material used to prepare a certain layer structure. For example, when preparing an aluminum nitride (AlN) layer containing p-type impurities, the precursor can be diethylberyllium, trimethylaluminum, and ammonia.

[0123] Heterojunction: A junction formed by the contact of two different semiconductor materials with different lattice constants, which results in lattice mismatch.

[0124] Interface states: In heterojunctions, in addition to the interface states generated at the interface between two different semiconductor materials due to lattice mismatch, interface states can also be generated due to the mismatch in the thermal expansion coefficients of the two materials, which causes interface distortion.

[0125] Semiconductor heterojunction energy band diagram: This diagram shows the energy changes of the lowest conduction band and the highest valence band on both sides of the heterojunction interface.

[0126] Donor and Acceptor: In doped semiconductors, most electrons and holes are provided by impurities. Impurities that can provide electrons are called donors, and impurities that can provide holes are called acceptors.

[0127] This application provides an electronic device that may include communication devices (e.g., base stations, mobile phones), wireless charging devices, medical devices, radar, navigation devices, radio frequency (RF) plasma lighting devices, RF induction and microwave heating devices, etc. This application does not impose any special limitations on the specific form of the aforementioned electronic device.

[0128] The aforementioned electronic devices almost all include radio frequency semiconductor devices, such as power amplifiers (PAs). The main function of a PA is to amplify radio frequency signals. Taking a base station as an example... Figure 2 A simplified structural diagram of a base station is provided. This base station includes a control unit, which comprises a radio transceiver, antennas, and related signal processing circuits. The control unit mainly consists of four components: a cell controller, a voice channel controller, a signaling channel controller, and a multiplexer interface for expansion. The base station control unit typically controls several base station transceivers. Through remote commands from the transceivers and mobile stations, the base station control unit is responsible for all mobile communication interface management, primarily the allocation, release, and management of radio channels.

[0129] Continue to combine Figure 2 The base station also includes a transmission unit, which is connected to the core network. Control signaling, voice calls, or data service information from the core network side are sent to the base station's control unit through the transmission unit, and the control unit processes these services.

[0130] Combined Figure 2 The base station also includes a baseband unit and a radio frequency (RF) unit. The baseband unit mainly performs functions such as baseband modulation and demodulation, allocation of radio resources, call processing, power control, and soft handover. The RF unit mainly performs the conversion between the air radio frequency channel and the baseband digital channel, then amplifies the signal through a power amplifier (PA), and then sends it to the antenna for transmission via the RF feeder. Terminal devices, such as mobile phones and tablets, receive the radio waves transmitted by the antenna through the wireless channel and then demodulate their own signal.

[0131] Continue to combine Figure 2 The base station also includes a power supply unit, which can be used to supply power to structures such as the transmission unit, baseband unit, and control unit.

[0132] Figure 3 A structural diagram of another electronic device is provided, taking a mobile phone as an example. The mobile phone may include a mid-frame 11, a back cover 12, and a display screen 13. The mid-frame 11 includes a carrier plate 111 for supporting the display screen 13, and a frame 110 around the carrier plate 111. The carrier plate 111 carries an RF unit and a PA device. The PA device amplifies the signal output by the RF unit and feeds it to the antenna in the mobile phone (for example, the antenna may be set along the edge of the frame 110) to transmit and receive signals.

[0133] With the development from fourth-generation wireless communication technologies (4G) to fifth-generation wireless communication technologies (5G), the functional requirements for the aforementioned radio frequency semiconductor devices are becoming increasingly higher, such as higher frequency, higher voltage, higher output power, and higher efficiency.

[0134] Among the available semiconductor materials, gallium nitride (GaN) has become a key material for fabricating radio frequency semiconductor devices due to its high thermal conductivity, high breakdown field strength, and high saturation electron mobility. For example, high electron mobility transistors (HEMTs) based on gallium nitride (GaN) are fabricated from GaN epitaxial single-crystal thin films grown on a single-crystal substrate. The single-crystal substrate is typically made of materials such as sapphire, silicon carbide (SiC), or silicon (Si). For instance, when the substrate is made of silicon single-crystal material, the resulting HEMT can be called a gallium nitride-on-silicon (GaN-on-Si) HEMT device.

[0135] like Figure 4 As shown, the radio frequency semiconductor device 101 in the above-mentioned device is disposed on the substrate 103, and the radio frequency semiconductor device 101 is disposed on the substrate 103 through a first electrical connection structure (e.g., a metal layer) 102, so that the radio frequency semiconductor device 101 can be interconnected with other electronic devices on the substrate 103.

[0136] The substrate 103 is then disposed on the circuit board 100 via the second electrical connection structure 200. For example, the circuit board 1 can be a printed circuit board (PCB), and the second electrical connection structure 200 can be a ball grid array (BGA) or other electrical connection structure.

[0137] exist Figure 4 The radio frequency semiconductor device 101 shown may include, for example: Figure 5 The structure shown, Figure 5 This is a cross-sectional view of a radio frequency semiconductor device 101, which includes: a silicon substrate, an aluminum nitride (AlN) layer formed on the silicon substrate, a GaN layer formed on the AlN layer, an aluminum gallium nitride (AlGaN) layer formed on the GaN layer, and a source, a gate, and a drain formed above the AlGaN layer.

[0138] However, in Figure 5 In the radio frequency semiconductor device 101 shown, parasitic conductive channels are easily generated on the surface of the silicon substrate near the AlN layer. The formation of these parasitic conductive channels is mainly due to two reasons: 1) The reaction chamber used to fabricate the radio frequency semiconductor device 101 contains Al and Ga residues, and after the AlN layer is produced, decomposition products of Al and Ga-containing precursors during subsequent high-temperature growth diffuse into the silicon substrate, thereby forming diffusion doping on the surface of the silicon substrate, such as... Figure 6 The P-type parasitic channel shown; 2) Due to the difference in material properties between AlN and Si, the energy bands of the layer structures of the two different materials are different, and high-conductivity-loss N-type parasitic channels are easily formed on the surface of the silicon substrate.

[0139] Whether it is Figure 6 As shown, whether a p-type parasitic channel is formed on the surface of a silicon substrate or an N-type parasitic channel is formed on the surface of a silicon substrate, the silicon substrate will become conductive, resulting in radio frequency loss and reducing the operating performance of the radio frequency semiconductor device.

[0140] Therefore, fabricating gallium nitride-on-silicon (GaN-on-Si) semiconductor devices with low radio frequency loss is key to realizing high-performance radio frequency device applications.

[0141] Based on this, embodiments of this application provide some achievable radio frequency semiconductor devices. These radio frequency semiconductor devices can suppress the formation of both P-type and N-type parasitic channels, thereby reducing the conduction loss of the radio frequency semiconductor device during operation, i.e., reducing radio frequency loss, and improving the output power, power gain, and efficiency of the radio frequency semiconductor device during operation.

[0142] The radio frequency semiconductor devices involved in the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0143] Combination Figure 7 , Figure 7 This is a cross-sectional view of a radio frequency semiconductor device 101 provided in this application. The radio frequency semiconductor device 101 includes: a substrate 1, a diffusion barrier layer 2 formed on the substrate 1, a nucleation layer 31 formed on the diffusion barrier layer 2, a channel layer 32 formed on the nucleation layer 31, and a barrier layer 33 formed on the channel layer 32. That is, the film structures of the diffusion barrier layer 2, the nucleation layer 31, the channel layer 32, and the barrier layer 33 are arranged along... Figure 7 The first direction L is shown, and the substrate 1 is stacked sequentially.

[0144] Additionally, please refer to... Figure 7The radio frequency semiconductor device 101 also includes a source 41, a gate 42, and a drain 43. The source 41, gate 42, and drain 43 are formed on the side of the barrier layer 33 away from the substrate 1.

[0145] Among them, Figure 7 The example shown includes a set of electrodes, comprising a source 41, a gate 42, and a drain 43. In other possible configurations, such as... Figure 8 As shown, Figure 8 An exemplary embodiment of the radio frequency semiconductor device 101 includes multiple sets of electrodes, for example, three sets of electrodes. The source electrodes 41 of the multiple sets of electrodes are electrically connected to each other, the gate electrodes 42 of the multiple sets of electrodes are electrically connected to each other, and the drain electrodes 43 of the multiple sets of electrodes are electrically connected to each other.

[0146] Furthermore, all of these electrodes extend in the same direction, and these electrodes can be arranged alternately at intervals along a direction perpendicular to their own extension direction.

[0147] The aforementioned substrate 1 may contain silicon, and a silicon-containing substrate 1 can be interpreted as follows: Figure 7 The surface A of substrate 1 opposite to diffusion barrier layer 2 can be a silicon surface. For example, in some embodiments, substrate 1 can be a silicon wafer, and surface A can correspond to the surface of the silicon wafer. In other embodiments, surface A can correspond to the silicon surface of a composite substrate (e.g., including a silicon layer and one or more lower layers disposed below the silicon layer). In other embodiments, surface A can correspond to the surface of a silicon portion of a silicon-on-insulator substrate. In still other embodiments, surface A can correspond to the surface of a silicon-on-sapphire substrate. In yet another embodiment, surface A can correspond to the silicon surface of an oxygen-implanted isolation substrate.

[0148] In addition, substrate 1 can be a high-resistivity substrate containing silicon. For example, the high-resistivity substrate containing silicon can be a substrate structure with a resistivity greater than or equal to 1000 Ω·cm, or a substrate structure with a resistivity greater than or equal to 2000 Ω·cm, or a substrate structure with a resistivity greater than or equal to 5000 Ω·cm.

[0149] Using the aforementioned silicon-containing high-resistivity substrates in devices operating at high frequencies (such as RF devices) can improve the performance of radio frequency semiconductor devices. For example, high resistivity can reduce substrate loss and thus reduce the power consumption of radio frequency semiconductor devices.

[0150] Continue reading Figure 7In this radio frequency semiconductor device 101, the nucleation layer 31, the channel layer 32, and the barrier layer 33, which are disposed along the first direction, i.e., the stacking direction L, can all contain group IIIA nitrides, that is, they can be made of any group IIIA nitride material. For example, group IIIA nitride materials can include one or a combination of at least two of boron nitride (BN), aluminum nitride (AlN), gallium nitride (GaN), indium nitride (InN), thallium nitride (TIN), and aluminum gallium nitride (AlGaN), or can include any alloy formed of group IIIA elements and group VA elements, such as indium gallium nitride (InGaN), aluminum indium gallium nitride (AlInGaN), gallium arsenide phosphide (GaAsPbN), and aluminum indium gallium arsenide phosphide (AlInGaAsPbN).

[0151] For example, nucleation layer 31 may contain aluminum nitride (AlN) material. Aluminum nitride (AlN) material can be understood as aluminum nitride (AlN) and any alloy thereof, such as aluminum gallium nitride (AlGaN), aluminum indium nitride (AlInN), aluminum indium gallium nitride (AlInGaN), aluminum indium gallium arsenide (AlInGaAsPbN), etc. These aluminum nitride (AlN) alloys have a high aluminum concentration and may contain small amounts of at least one of gallium and indium, or may not contain gallium and indium, meaning the alloy concentration is relatively low compared to the aluminum concentration.

[0152] Continue reading Figure 7 In this radio frequency semiconductor device 101, a diffusion barrier layer 2 is formed between the substrate 1 and the nucleation layer 31 along the stacking direction P. The diffusion barrier layer 2 can be formed on the surface A of the substrate 1, that is, it can directly cover the surface A of the substrate 1.

[0153] The diffusion barrier layer 2 can be made of a variety of materials, such as metals (e.g., Al), or one or a combination of at least two of silicon oxides, silicon nitrides, aluminum oxides, aluminum silicon compounds, aluminum silicon nitrides, etc.

[0154] In some process flows, for example, it may be difficult to directly perform heteroepitaxial growth of the nucleation layer 31, which includes AlN, on the substrate 1. By first forming a diffusion barrier layer 2 containing aluminum on the substrate 1, the lattice difference between the group IIIA nitride material of the nucleation layer 31 and the lattice of the substrate 1 can be mitigated. Alternatively, the diffusion barrier layer 2 can be understood as mitigating the lattice mismatch between the substrate 1 and aluminum nitride, thereby making the aluminum nitride lattice grow more uniformly.

[0155] In addition, when the nucleation layer 31 is formed on the diffusion barrier layer 2, the diffusion barrier layer 2 can also suppress the diffusion of group IIIA elements (e.g., Al) in the nucleation layer 31 into the silicon-containing substrate 1. Alternatively, when the channel layer 32 or the barrier layer 33 is formed on the nucleation layer 31, the diffusion barrier layer 2 can also suppress the diffusion of group IIIA elements (e.g., Al and Ga) in the channel layer 32 or the barrier layer 33 into the silicon-containing substrate 1. In this way, the concentration of these group IIIA elements diffusing into the substrate 1 can be reduced, thereby suppressing the formation of P-type parasitic channels in the substrate 1 by these group IIIA elements diffusing into the substrate 1, thus weakening the conductivity of the substrate 1.

[0156] When forming the diffusion barrier layer 2 on the substrate 1, various process methods can be used to obtain the diffusion barrier layer 2. For example, ammonia gas can be used for pre-nitriding to form a silicon nitride thin film as the diffusion barrier layer 2 on the silicon-containing substrate 1; as another example, an alumina thin film can be formed on the silicon-containing substrate 1 as the diffusion barrier layer 2 using epitaxial growth technology, thereby reducing the conductivity of the substrate 1 after epitaxial growth.

[0157] Based on the above description of the diffusion barrier layer 2, it can be understood that: the diffusion barrier layer 2 not only alleviates the lattice mismatch between the substrate 1 and the nucleation layer 31, making the lattice growth of the nucleation layer 31 more orderly, but also enables the subsequent formation of the channel layer 32 or barrier layer 33 to have relatively orderly lattices and low dislocation densities, thereby improving the output power, power gain, and efficiency of the RF semiconductor device. Furthermore, the diffusion barrier layer 2 also suppresses the diffusion of Group IIIA elements from the nucleation layer 31, channel layer 32, or barrier layer 33 into the substrate 1, thus suppressing the formation of P-type parasitic channels in the substrate 1.

[0158] Figure 9 This is a structural diagram of another radio frequency semiconductor device 101 provided in the embodiments of this application, and the above... Figure 7 Compared to the radio frequency semiconductor device 101 shown, Figure 9 The radio frequency semiconductor device 101 shown includes not only a substrate 1, a diffusion barrier layer 2, a nucleation layer 31, a channel layer 32, and a barrier layer 33, but also other film structures. For example, see [link to relevant documentation]. Figure 9 The radio frequency semiconductor device 101 may also include a stress buffer layer 34, which may be formed on the side of the nucleation layer 31 away from the diffusion barrier layer 2, that is, on the nucleation layer 31.

[0159] The stress buffer layer 34 includes, but is not limited to, single-layer or multi-layer group IIIA nitrides and multi-component group IIIA nitrides, or superlattice structures composed of them.

[0160] In some alternative embodiments, the stress buffer layer 34 may comprise a compositionally graded Group IIIA nitride material, for example, the stress buffer layer 34 may comprise AlxGa(1-x)N, wherein x may be along the growth direction (i.e., Figure 9 (As shown, x gradually decreases along the direction away from substrate 1; for example, x can decrease from a value of 1 to a value of 0.)

[0161] The composition-gradient stress buffer layer 34 can be gradually varied according to the following exemplary pattern.

[0162] For example, the value of x in the composition of AlxGa(1-x)N gradually changes from 1 at the lower surface B of the stress buffer layer 34 to 0 at the upper surface C of the stress buffer layer 34.

[0163] For example, in the AlxGa(1-x)N composition, the value of x discontinuously changes from 1 at the lower surface B of the stress buffer layer 34 to 0 at the upper surface C of the stress buffer layer 34. For instance, from the lower surface B of the stress buffer layer 34 to the upper surface C of the stress buffer layer 34, the values ​​of x, x, and x change discontinuously from AlN to AlN. 0.6 Ga 0.4 N and Al 0.4 Ga 0.6 N, and Al 0.2 Ga 0.8 N undergoes a discontinuous gradual change.

[0164] For example, the value of x in the AlxGa(1-x)N composition changes parabolically from 1 at the lower surface B of the stress buffer layer 34 to 0 at the upper surface C of the stress buffer layer 34; or it changes in other continuous or discontinuous gradual forms.

[0165] Additionally, in some alternative embodiments, gallium nitride alloys such as AlxInyGa(1-xy)N and InyGa(1-y)N can be selected for formation. The concentration of at least one of the alloying elements, such as Ga, Al, and In, varies along the thickness of the stress buffer layer. For example, when the stress buffer layer has an AlxInyGa(1-y)N composition, at least one of x and y can be varied. As another example, in some embodiments of the InyGa(1-y)N composition of the stress buffer layer, y can be varied.

[0166] Continue to combine Figure 9 The channel layer 32 of the radio frequency semiconductor device 101 can be formed on the side of the stress buffer layer 34 away from the nucleation layer 31, that is, formed on top of the stress buffer layer 34. The channel layer 32 can contain gallium nitride (GaN) material. Figure 9When the upper surface of the stress buffer layer 34 is made of GaN material and the channel layer 32 is also made of GaN material, the channel layer 32 made of GaN material is generally doped (for example, doped with elements such as carbon or iron) to form a high-resistivity GaN layer, that is, the purpose of suppressing leakage current and increasing breakdown voltage is achieved by doping. However, the channel layer 32 made of GaN material is used for current conduction, which requires a high-quality GaN epitaxial layer with low impurities.

[0167] The channel layer 32 may comprise a single channel layer, or may comprise at least two channel layers stacked one on top of another.

[0168] With continued reference to Figure 9 , the radio frequency semiconductor device 101 may further comprise an isolation layer 35, and the isolation layer 35 may be formed on a side of the channel layer 32 away from the stress buffer layer 34, that is, formed over the channel layer 32.

[0169] The isolation layer 35 may comprise an AlN material. The polarization effect of the AlGaN, AlN and GaN structure can generate a higher two-dimensional electron gas concentration, and the probability of the two-dimensional electron gas penetrating into the barrier layer is reduced. The isolation layer 35 can reduce alloy disorder scattering so as to improve mobility, which is beneficial to improving the output characteristics of the device.

[0170] The thickness of the isolation layer 35 may be less than or equal to 2 nm, and greater than or equal to 0.1 nm.

[0171] With continued reference to Figure 9 , the barrier layer 33 in the radio frequency semiconductor device 101 may be formed on a side of the isolation layer 35 away from the channel layer 32, that is, formed over the isolation layer 35.

[0172] The barrier layer 33 may include, but is not limited to, group IIIA nitrides and multi-component group IIIA nitrides, such as Al a Ga b In 1-a-b N, where 0≤a+b≤1. In some other specific embodiments, Al may also be selected z Ga 1-z N, where 0.1<z<0.3. In some other specific embodiments, In may also be selected 0.17 Al 0.83 N.

[0173] Herein, the aluminum content in the barrier layer 33 is different from the aluminum content in the stress buffer layer 34 and the channel layer 32.

[0174] The thickness of the barrier layer 33 may be less than or equal to 100 nm, and greater than or equal to 5 nm, for example, may be selected to be less than or equal to 25 nm, and greater than or equal to 10 nm.

[0175] The barrier layer 33 is used in conjunction with the channel layer 32 to generate a two-dimensional electron gas (2DEG) channel region between the channel layer 32 and the barrier layer 33 through polarization, providing a channel for rapid electron transport and thus conducting current.

[0176] Continue to combine Figure 9 The radio frequency semiconductor device 101 may also include a cap layer 36, which may be formed on the side of the barrier layer 33 away from the isolation layer 35, that is, on the barrier layer 33.

[0177] In some alternative embodiments, with the capping layer 36, the resulting surface morphology can be smoother and have fewer surface defects compared to a surface formed without the capping layer. Furthermore, using GaN material to terminate the epitaxial layer growth also facilitates subsequent chemical treatment of the epitaxial layer surface.

[0178] The cap layer 36 may include, but is not limited to, a grown nitride or oxide layer. For example, in some embodiments, a single-crystal GaN layer may be selected for growth, and in other embodiments, a polycrystalline SiNx layer may be selected for growth.

[0179] In addition, the cap layer 36 can be monocrystalline, polycrystalline, or amorphous.

[0180] Among the available process options, metal-organic vapor phase epitaxy (MOVPE), also known as metal-organic chemical vapor deposition (MOCVD), can be used as the growth technique to form a group IIIA nitride formation layer on the diffusion barrier layer 2, comprising a nucleation layer 31, a channel layer 32, a barrier layer 33, a stress buffer layer 34, an isolation layer 35, and a cap layer 36. Alternatively, molecular beam epitaxy (MBE) can be used as the growth technique to form a group IIIA nitride formation layer on the diffusion barrier layer 2, comprising a nucleation layer 31, a channel layer 32, and a barrier layer 33. Of course, other epitaxial growth techniques can also be used.

[0181] In some processes, such as when growing nucleation layer 31 using MOVPE technology, a large amount of NH3 needs to be introduced. AlN layers grown in an NH3 atmosphere usually exhibit n-type conductivity due to the presence of defects. In addition, since the lattice constant of AlN is smaller than that of Si, the AlN layer will be subjected to greater tensile stress, resulting in a stronger polarization effect, which causes the energy band of AlN to bend.

[0182] The following is through Figures 10a to 10c The semiconductor band diagram shown explains some phenomena that occur when the nucleation layer 31 exhibits n-type conductivity.

[0183] Figure 10a The diagrams shown are the band structures of AlN and Si. Figure 10b The diagrams shown are the band structures of AlN and Si, which exhibit n-type conductivity. Figure 10c The diagram shows the band structure of a heterojunction formed by AlN and Si, which exhibits n-type conductivity.

[0184] Among them, Figures 10a to 10c In the band structure diagram, the physical meanings of the various quantities are as follows:

[0185] E0: Vacuum energy level. Represents the minimum energy an electron must possess to escape a semiconductor and enter a vacuum.

[0186] E F Fermi level. For a microscopic system composed of fermions (which can be electrons, protons, or neutrons), each fermion exists in its own quantum energy state. In band theory, the Fermi level can be considered as a hypothetical energy level that an electron has a 50% probability of occupying at thermodynamic equilibrium.

[0187] E C : This indicates the bottom of the guide strip with the band width not shown.

[0188] E V : This indicates the top of the price band, representing the width of the no-bandwidth area.

[0189] For example: Band gap. This is the energy difference between the valence band top and the conduction band bottom, which is the minimum average energy required to generate intrinsic excitations. For instance, Figure 10a In the diagram, Eg-AlN represents the bandgap of AlN, and Eg-Si represents the bandgap of Si. The size of the bandgap determines whether a material exhibits semiconductor or insulating properties. For example, Eg-AlN can have a bandgap of 6.2 eV, while Eg-Si can have a bandgap of 1.1 eV. This means that Si has a smaller bandgap, so when the temperature rises, electrons are easily transferred to the conduction band of Si, making Si conductive. However, AlN has a larger bandgap, and compared to Si, it remains a poor conductor even at higher temperatures. Therefore, in semiconductor devices, when Si is used as the substrate material, controlling electron migration into Si can effectively control the conductivity of Si.

[0190] X: Electron affinity. This is the energy required for an electron to move from the bottom of the conduction band to a vacuum level. For example, Figure 10aIn the diagram, XAlN represents the electron affinity of AlN, and XSi represents the electron affinity of Si. For example, XAlN can be 1.15 eV, and XSi can be 4.05 eV.

[0191] contrast Figure 10a and Figure 10b When the nucleation layer 31 includes AlN exhibiting n-type conductivity, such as Figure 10b As shown, the AlN layer will be subjected to greater tensile stress, resulting in a stronger polarization effect, which causes band bending. The Fermi level of the donor AlN is closer to the conduction band bottom than the Fermi level of Si.

[0192] like Figure 10b As shown, when two different semiconductors (AlN and Si, n-type conductivity types) are brought into close contact to form a heterojunction, electron transfer will inevitably occur to achieve equilibrium until the Fermi levels throughout the system are completely identical. In other words, electrons from the donor n-type AlN will transfer to Si. This electron transfer will cause changes in the energy bands near the interface (including the vacuum level E0 and the conduction band bottom E0). C Price band top E V ,like Figure 10c The vacuum energy level E0 in the middle, and Figure 10b and Figure 10c E in the conduction band bottom C Price band top E V The bending of the conduction band causes the local conduction band space to be at the Fermi level E. F Electrons will spontaneously move and accumulate in this region. In this way, the band structure between n-type AlN and Si causes an energy barrier to form at the interface, as shown in the image. Figure 10c The N-type parasitic channel of the inversion layer shown typically has high conductivity, which increases the RF loss of the RF semiconductor device and thus increases its power consumption.

[0193] In order to overcome the above-mentioned problems Figure 10c The N-type parasitic channels of the inversion layer shown are as follows: Figure 11 As shown in the embodiments of this application, another radio frequency semiconductor device is provided. Figure 11 This is a schematic diagram of the structure of the radio frequency semiconductor device 101. The radio frequency semiconductor device 101 and the above-mentioned... Figure 7 The radio frequency semiconductor device 101 shown is similar in that it also includes a substrate 1, a diffusion barrier layer 2, a nucleation layer 31, a channel layer 32, a barrier layer 33, and electrodes. (And the above...) Figure 7 The difference in the radio frequency semiconductor device 101 shown is that: Figure 11 In the structure shown, the nucleation layer 31 contains P-type impurities.

[0194] The P-type impurity may contain at least one of Group II elements. For example, it may be one or a combination of at least two of Group IIA elements: beryllium (Be), magnesium (Mg), calcium (Ca), strontium (Sr), and barium (Ba).

[0195] As another example, one or a combination of at least two of the group IIB elements zinc (Zn), mercury (Hg), and cadmium (Cd) can be used.

[0196] As another example, a combination of at least two of beryllium (Be), magnesium (Mg), calcium (Ca), strontium (Sr), barium (Ba), zinc (Zn), mercury (Hg), and cadmium (Cd) may be used.

[0197] As another example, the P-type impurity may also contain at least one of the elements in IVA, such as at least one of carbon (C) and silicon (Si).

[0198] Alternatively, the P-type impurity may also include a combination of at least two of the different elements shown above. For example, it may include magnesium (Mg) and carbon (C); or, for another example, zinc (Zn) and silicon (Si).

[0199] exist Figure 11 In the radio frequency semiconductor device 101 shown, the thickness of the diffusion barrier layer 2 along the stacking direction can be less than 1 nm. Therefore, the diffusion barrier layer 2 can prevent the diffusion of Group IIIA ions and / or Group IIIA atoms from the nucleation layer 31, the channel layer 32, and the barrier layer 33 into the substrate 1. However, since the diffusion barrier layer 2 is relatively thin, it cannot confine electrons, that is, it cannot prevent electrons in the nucleation layer 31 from migrating into the substrate 1. In this case, it can be considered that a heterojunction is formed between the nucleation layer 31 and the substrate 1.

[0200] Figure 12a The diagrams shown are the band structures of AlN and Si, which exhibit p-type conductivity. Figure 12b The diagram shows the band structure of a heterojunction formed by AlN and Si, which is a P-type conductivity type. The physical meaning of each quantity in the band structure has already been explained above and will not be repeated here.

[0201] See Figure 12a When AlN exhibits p-type conductivity, the Fermi level of the acceptor AlN is closer to the valence band top than that of Si. Similarly, when p-type conductive AlN and Si form a heterojunction through close contact, the Fermi level E0 after band alignment... F It is within the band gap, meaning there is no region where electrons can spontaneously accumulate. Therefore, in comparison... Figure 12b and Figure 10cIt can be seen that when the nucleation layer 31 is P-type, it effectively suppresses the formation of parasitic channels in the inversion layer, reduces the RF power consumption of the RF semiconductor device, and improves the operating performance of the RF semiconductor device.

[0202] See you later Figure 12a Due to the E of the heterojunction formed by AlN and Si F Within the restricted zone, both N-type and P-type parasitic channels are effectively suppressed.

[0203] See also Figure 11 The radio frequency semiconductor device 101 provided in this application embodiment not only includes a diffusion barrier layer 2 formed between the nucleation layer 31 and the substrate 1, but also has P-type impurities in the nucleation layer 31. This design allows the diffusion barrier layer 2 to suppress the diffusion of group IIIA atoms (e.g., aluminum and gallium) into the substrate 1, suppressing the formation of P-type parasitic channels in the substrate 1. The diffusion barrier layer 2 also alleviates lattice mismatch between the nucleation layer 31 and the substrate 1. Furthermore, the P-type impurities formed in the nucleation layer 31 suppress the formation of inversion layer N-type parasitic channels on the surface of the substrate 1. In summary, when the radio frequency semiconductor device 101 includes a diffusion barrier layer 2 and a nucleation layer 31 containing P-type impurities, it not only suppresses the formation of P-type parasitic channels in the substrate 1 but also suppresses the formation of N-type parasitic channels in the substrate 1, thereby effectively suppressing the conductivity of the substrate 1, reducing parasitic losses in the substrate 1, and improving the performance of the radio frequency semiconductor device 101.

[0204] Even in some embodiments, a small number of electrons migrate from the nucleation layer 31 to the substrate 1 to form a P parasitic channel with weak conductivity. However, by forming P-type impurities in the nucleation layer 31, the conductivity of the substrate 1 is still very weak, which will correspondingly reduce the conductivity of the substrate 1 and reduce the radio frequency power consumption of the substrate 1.

[0205] In some optional implementation architectures, Figure 11 The concentration ρ of P-type impurities in nucleation layer 31 can be: 1×10 15 cm -3 ≤ρ≤1×10 20 cm -3 For example, when Mg is chosen as the P-type impurity, the concentration ρ of Mg can be: 1 × 10⁻⁶ 15 cm -3 ≤ρ≤1×10 20 cm -3 For example, the concentration of Mg can be chosen to be 1×10⁻⁶. 18 cm -3 For example, when the p-type impurity is a mixture of C and Si, the concentrations of both C and Si can be 1 × 10⁻⁶. 18 cm -3 .

[0206] Furthermore, in some feasible structures, the thickness d of the nucleation layer 31 containing P-type impurities ranges from 1 nm to 300 nm. For example, d = 150 nm, d = 200 nm, d = 250 nm, or d = 300 nm can be chosen. Here, the thickness d of the nucleation layer 31 refers to the thickness of the nucleation layer 31 along the stacking direction of these film structures, that is, along... Figure 9 The thickness dimension in the L direction is shown.

[0207] The radio frequency semiconductor device 101 provided in the embodiments of this application, such as Figure 11 As shown, the RF loss of substrate 1 or other epitaxial layer structures can be measured by fabricating coplanar waveguides (CPWs) on the surface of the epitaxial layer. For example, it is possible to... Figure 11 The surface of the barrier layer 33 is fabricated with a coplanar waveguide (CPW) for measurement. In some embodiments, by measuring with CPW, the RF loss of the entire RF device on a gallium nitride epitaxial substrate on silicon can be less than 0.25 dB / mm in the frequency band below 40 GHz. Compared with the existing loss greater than 0.4 dB / mm, the RF loss of the RF semiconductor device given in this application embodiment is significantly reduced.

[0208] The full width at half maximum (FWHM) of the rocking curve of the nucleation layer 31 containing P-type impurities can be measured by X-ray diffraction (XRD). The aforementioned FWHM refers to the measurement of the diffraction peaks in omega scanning mode of the 0002 crystal plane of the AlN nucleation layer. In this embodiment, the FWHM of the rocking curve of the 0002 crystal plane of the nucleation layer 31 containing P-type impurities (for example, if the AlN in the nucleation layer 31 is hexagonal, the 0002 crystal plane is the crystal plane parallel to the surface of the stack) is less than 800 arcseconds.

[0209] The surface roughness of the nucleation layer 31 containing P-type impurities can be measured using atomic force microscopy (AFM). In this embodiment, the surface roughness of the nucleation layer 31 can be less than 1 nm.

[0210] To further suppress the radio frequency loss of radio frequency semiconductor devices, such as Figure 13 This application also provides a structure for a radio frequency semiconductor device 101, and Figure 13 This is a structural diagram of the provided radio frequency semiconductor device 101. Figure 13 The structure shown and the above Figure 11The difference in the structure shown is that the nucleation layer 31 includes a first nucleation layer 311 and a second nucleation layer 312 stacked along the first direction L. Specifically, the first nucleation layer 311 can be stacked on the diffusion barrier layer 2, and the second nucleation layer 312 can be stacked on the first nucleation layer 311. Furthermore, P-type impurities are formed in both the first nucleation layer 311 and the second nucleation layer 312.

[0211] Figure 13 An exemplary two-layer structure is given, comprising a stacked first nucleating layer 311 and a second nucleating layer 312. In some other implementable structures, [the structure can be...]. Figure 13 Based on the structure shown, there are more nucleation layers, such as three, four, or even more layers.

[0212] about Figure 13 In the example structure, the materials selected for the P-type impurities in the first nucleation layer 311 and the second nucleation layer 312 can refer to the above, and the specific materials of the P-type impurities will not be described here.

[0213] In some feasible structures, such as Figure 13 The P-type impurity elements in the first nucleation layer 311 and the P-type impurity elements in the second nucleation layer 312 can be different.

[0214] The difference between the P-type impurity elements in the first nucleation layer 311 and those in the second nucleation layer 312 can be explained as follows:

[0215] For example, when either the first nucleation layer 311 or the second nucleation layer 312 contains only one type of P-type impurity element, the P-type impurity elements in the two nucleation layers are different. That is, the P-type impurity element in the first nucleation layer 311 is considered to be different from the P-type impurity element in the second nucleation layer 312. For example, the P-type impurity element in the first nucleation layer 311 is magnesium (Mg), and the P-type impurity element in the second nucleation layer 312 is carbon (C).

[0216] For example, when both the first nucleation layer 311 and the second nucleation layer 312 include at least two types of P-type impurity elements, the P-type impurity elements in the first nucleation layer 311 are different from those in the second nucleation layer 312. For instance, if the P-type impurity elements in the first nucleation layer 311 are magnesium (Mg) and carbon (C), then the P-type impurity elements in the second nucleation layer 312 are calcium (Ca) and silicon (Si), which are different from both magnesium (Mg) and carbon (C).

[0217] For example, when both the first nucleation layer 311 and the second nucleation layer 312 include at least two types of P-type impurity elements, the P-type impurity elements in the first nucleation layer 311 and the P-type impurity elements in the second nucleation layer 312 are considered to be partially the same and partially different. For instance, if the P-type impurity elements in the first nucleation layer 311 are magnesium (Mg) and carbon (C), then if the P-type impurity elements in the second nucleation layer 312 are magnesium (Mg) and silicon (Si), then the P-type impurity elements in the first nucleation layer 311 are different from those in the second nucleation layer 312.

[0218] like Figure 14 This is a structural diagram of another radio frequency semiconductor device 101 provided in an embodiment of this application. See Figure 14 As shown, in this radio frequency semiconductor device 101, the nucleation layer 31 includes a stacked first nucleation layer 311 and a second nucleation layer 312, and along... Figure 14 As shown along the stacking direction L, the thickness d1 of the first nucleating layer 311 is not equal to the thickness d2 of the second nucleating layer 312. For example, the thickness d1 of the first nucleating layer 311 can be less than the thickness d2 of the second nucleating layer 312, or the thickness d1 of the first nucleating layer 311 can be greater than the thickness d2 of the second nucleating layer 312.

[0219] In some examples, the thickness d1 of the first nucleation layer 311 can be less than or equal to 100 nm and greater than or equal to 1 nm, for example, it can be 50 nm.

[0220] In other examples, the thickness d2 of the second nucleation layer 312 can be less than or equal to 250 nm and greater than or equal to 100 nm, for example, it can be 150 nm.

[0221] When the nucleation layer 31 comprises a multilayer structure, in order to prevent P-type impurities from forming parasitic channels within the nucleation layer 31 itself, the concentration of P-type impurities in the multilayer nucleation layer 31 can be made different, for example, see... Figure 15 As shown, Figure 15 This is yet another structural diagram of a radio frequency semiconductor device 101 provided in the embodiments of this application. Figure 15The concentration ρ1 of P-type impurities in the first nucleation layer 311 is greater than the concentration ρ2 of P-type impurities in the second nucleation layer 312, and the first nucleation layer 311 is positioned closer to the diffusion barrier layer 2 than the second nucleation layer 312. The greater concentration of P-type impurities in the first nucleation layer 311 than in the second nucleation layer 312 can be understood as follows: the concentration of P-type impurities in the first nucleation layer 311 differs from that in the second nucleation layer 312 by at least one order of magnitude, i.e., ρ1 / ρ2 > 10. For example, the concentration ρ1 of P-type impurities in the first nucleation layer 311 is 10. 19 cm -3 Then the concentration ρ2 of the P-type impurity in the second nucleation layer 312 can be 10. 17 cm -3 .

[0222] exist Figure 15 In the diagram, black dots represent P-type impurities. A higher number of P-type impurities indicates a higher concentration of P-type impurities, while a lower number of P-type impurities indicates a lower concentration of P-type impurities.

[0223] In some examples, the concentration ρ1 of the P-type impurities in the first nucleation layer 311 can be: 1 × 10⁻⁶ 17 cm -3 ≤ρ≤1×10 20 cm -3 For example, the concentration ρ1 of a P-type impurity can be 1 × 10⁻⁶. 19 cm -3 .

[0224] In other examples, the concentration ρ2 of the P-type impurities in the second nucleation layer 312 can be: 1 × 10⁻⁶. 15 cm -3 ≤ρ≤1×10 18 cm -3 For example, the concentration ρ2 of P-type impurities can be 1×10⁻⁶. 17 cm -3 .

[0225] With this design, the first nucleation layer 311, with a higher concentration of P-type impurities, can effectively suppress the formation of inversion layer N-type parasitic channels on the surface of substrate 1, reducing the conductivity of substrate 1 and thus lowering the RF power consumption of the RF semiconductor device. Simultaneously, the second nucleation layer 312, with a lower concentration of P-type impurities, can prevent excessively high P-type impurity concentrations from causing parasitic conductance within the nucleation layer itself. In some feasible structures, Figure 15 The radio frequency semiconductor device shown can achieve a radio frequency conduction loss of less than 0.20 dB / mm in the frequency band below 40 GHz, which can significantly further reduce the radio frequency loss of the radio frequency device.

[0226] In some embodiments, if in Figure 15 Based on this, when more nucleation layers of phase stacking are set, the concentration of P-type impurities in the nucleation layers gradually decreases along the direction away from the substrate 1. For example, it can be reduced until the concentration of P-type impurities in the topmost nucleation layer away from the substrate 1 is close to zero.

[0227] In some feasible structures, variations in the concentration of P-type impurities in the multilayer nucleation layers can be combined with variations in the thickness of the multilayer nucleation layers to form other radio frequency semiconductor devices. For example, see... Figure 16 As shown, Figure 16 The diagram shown is a structural diagram of another radio frequency semiconductor device 101 provided in this application embodiment. In this embodiment, the thickness of the multilayer nucleation layer gradually increases along the direction away from the substrate 1, and the concentration of P-type impurities in the multilayer nucleation layer gradually decreases, for example, Figure 16 The thickness of the first nucleation layer 311 near the substrate 1 is less than the thickness of the second nucleation layer 312 away from the substrate 1, and the concentration of P-type impurities in the first nucleation layer 311 is greater than the concentration of P-type impurities in the second nucleation layer 312.

[0228] In preparation Figure 16 In the illustrated embodiment, the first nucleation layer 311, located near the substrate 1, has a large lattice mismatch with the substrate 1 and is generally grown at low temperatures, resulting in poor crystal quality. Therefore, the thickness of the first nucleation layer 311 is kept as thin as possible. When growing the second nucleation layer 312, it is generally grown at high temperatures, and the crystal quality gradually improves with increasing thickness. Therefore, the thickness of the second nucleation layer 312 can be greater than the thickness of the first nucleation layer 311. Furthermore, the first nucleation layer 311 is highly doped compared to the second nucleation layer 312 to increase its conductivity. Thus, while ensuring sufficient suppression of N-type parasitic channel formation in the substrate 1, the thickness of the first nucleation layer is kept as thin as possible to reduce its own conductivity.

[0229] Figure 17 This is a structural diagram of another radio frequency semiconductor device provided in the embodiments of this application, and Figure 17 Only the substrate 1, diffusion barrier layer 2, first nucleation layer 311 and second nucleation layer 312 of the radio frequency semiconductor device are shown; other film structures are not shown.

[0230] exist Figure 17The diagram shows a crystal arrangement model of the first nucleation layer 311 and the second nucleation layer 312. The black circles represent the crystal arrangement in the first nucleation layer 311, and the white circles represent the crystal arrangement in the second nucleation layer 312. These two crystal model diagrams are merely illustrative and do not constitute a specific limitation on the crystal arrangement of the first nucleation layer 311 and the second nucleation layer 312 of this application.

[0231] Depend on Figure 17 As can be seen from the exemplary crystal arrangement of the first nucleation layer 311 and the second nucleation layer 312 in this embodiment, the dislocation density of the second nucleation layer 312 disposed away from the substrate 1 is less than the dislocation density of the first nucleation layer 311 disposed close to the substrate 1. That is, the dislocation density of the nucleation layer can gradually decrease along the direction away from the substrate 1.

[0232] In feasible processes, when fabricating multilayer nucleation layers, the dislocation density of the nucleation layer can be gradually reduced along the direction away from the substrate 1 by controlling the growth temperature. For example, in Figure 17 In this process, the growth temperature can be controlled at 700°C when forming the first nucleation layer 311, while the temperature for growing the second nucleation layer 312 is 1100°C. This allows the crystal arrangement in the nucleation layer to gradually become more orderly along the direction away from the substrate 1, which is beneficial for forming a channel layer 32, barrier layer 33, or other film structure with orderly crystal arrangement on the nucleation layer. In addition, it helps to avoid excessive tensile stress in the first nucleation layer 311, thereby reducing the risk of crack formation.

[0233] Combined Figures 13 to 17 In these feasible structures, not only are diffusion barrier layers 2 formed in the substrate 1 and nucleation layer 31, but also nucleation layer structures with at least two stacked phases are included. That is, by using diffusion barrier layer 2 to prevent the diffusion of group IIIA elements in nucleation layer 31 into substrate 1 and suppressing the formation of P-type parasitic channels in substrate 1, the nucleation layer near the bottom layer of the multilayer nucleation layer can also be used to suppress the formation of N-type parasitic channels in substrate 1. At the same time, these multilayer nucleation layers also lay the foundation for the high-quality formation of other film layer structures in subsequent processes, that is, it can make the crystal arrangement of channel layer 32 and barrier layer 33 more orderly, the electron mobility higher, and prevent the introduction of P-type parasitic channels in the nucleation layer itself, thus making the performance of the radio frequency semiconductor device better.

[0234] In addition to the various radio frequency semiconductor devices with different structures described above, this application also provides several different radio frequency semiconductor devices to suppress the formation of parasitic channels in substrate 1, as shown below.

[0235] Figure 18This is a structural diagram of a radio frequency semiconductor device provided in an embodiment of this application. The radio frequency semiconductor device 101 with this structure differs from the radio frequency semiconductor devices mentioned above in that it may not include the diffusion barrier layer 2, but instead the nucleation layer 31 can be formed on the substrate 1.

[0236] in addition, Figure 18 The nucleation layer 31 in the middle includes at least two stacked layers, for example, the nucleation layer 31 includes a first nucleation layer 311 and a second nucleation layer 312 stacked in phase.

[0237] exist Figure 18 In the radio frequency semiconductor device 101 shown, it is assumed that the nucleation layer 31, which includes the first nucleation layer 311 and the second nucleation layer 312, can be directly covered on the substrate 1.

[0238] Also, you can Figure 18 Based on this, more nucleation layers can be stacked, such as three, four, or even more layers.

[0239] When Figure 18 When the nucleation layer shown includes at least two layers, the materials, concentrations, dislocation densities, growth temperatures, etc., that can be selected for these multi-layer nucleation layers can all refer to the above. Figures 13 to 17 The structure shown describes the first nucleation layer 311 and the second nucleation layer 312.

[0240] Example, Figure 19 This is a structural diagram of a radio frequency semiconductor device including a first nucleation layer 311 and a second nucleation layer 312, according to an embodiment of this application. In this embodiment, the concentration ρ1 of P-type impurities in the first nucleation layer 311 is greater than the concentration ρ2 of P-type impurities in the second nucleation layer 312.

[0241] Continue reading Figure 19 The first nucleation layer 311 and the adjacent substrate 1 form a heterojunction due to their different energy bands. However, since the first nucleation layer 311 contains a high concentration of P-type impurities, the formation of an inversion layer N-type parasitic channel on the surface of the substrate 1 can be effectively suppressed, thereby reducing the conductivity of the substrate 1 and reducing the RF power consumption of the RF semiconductor device.

[0242] In addition, the second nucleation layer 312 with a lower concentration of P-type impurities can also prevent parasitic conductivity caused by excessively high concentrations of P-type impurities within the nucleation layer itself.

[0243] Another example, Figure 20 This is a structural diagram of another radio frequency semiconductor device including a nucleation layer, as provided in an embodiment of this application. Figure 20 and the above Figure 19 The difference in the structure shown is that: Figure 20The process includes not only the first nucleation layer 311 and the second nucleation layer 312 stacked on the phase, but also a third nucleation layer 313 stacked on the second nucleation layer 312, and the third nucleation layer 313 also contains P-type impurities.

[0244] Furthermore, along the L direction away from the substrate 1, the concentration of P-type impurities in the first nucleation layer 311, the second nucleation layer 312, and the third nucleation layer 313 gradually decreases.

[0245] Another example, Figure 21 This is a structural diagram of another radio frequency semiconductor device including a nucleation layer, as provided in the embodiments of this application. Figure 21 The structure shown and Figure 20 The structure shown is similar, also including a first nucleation layer 311, a second nucleation layer 312, and a third nucleation layer 313 stacked in phases, and the concentration of P-type impurities in the first nucleation layer 311, the second nucleation layer 312, and the third nucleation layer 313 gradually decreases along the L direction away from the substrate 1. Figure 20 In the structure shown, the first nucleation layer 311, the second nucleation layer 312, and the third nucleation layer 313 have essentially the same thickness, while... Figure 21 In the embodiment shown, the thickness of the first nucleation layer 311, the second nucleation layer 312, and the third nucleation layer 313 gradually increases along the L direction away from the substrate 1.

[0246] The above describes various radio frequency semiconductor devices with different structures provided in the embodiments of this application. The following describes some methods for fabricating radio frequency semiconductor devices, as detailed below.

[0247] Figure 22 This is a process flow diagram of the fabrication method for this radio frequency semiconductor device. Figures 23a to 23d This is a cross-sectional view showing the completion of each step in the manufacturing process of this radio frequency semiconductor device.

[0248] like Figure 22 Step S11 in the middle, and Figure 23a , Figure 23a Is execution Figure 22 The structural diagram after step S11.

[0249] like Figure 22 Step S11 and Figure 23a A first nucleation layer 311 is formed on a substrate 1 containing silicon. The first nucleation layer 311 includes a group IIIA nitride doped with a p-type impurity.

[0250] When the first nucleation layer 311 is formed, epitaxial growth technology can be used. For example, epitaxial growth technology includes, but is not limited to, metal-organic chemical vapor phase epitaxy (MOVPE), molecular beam epitaxy (MBE), or other single-crystal epitaxial growth technology.

[0251] Furthermore, when preparing the first nucleation layer 311, the corresponding P-type impurity doping precursor can include, but is not limited to, diethylberyllium (DEBe), magnesia-cerammonium (Cp2Mg), dimethylzinc (DMZn), calcium diisopropyl dithiocarbamate Ca(DPTC)2, methane (CH4), ethylene (C2H4), silane (SiH4), or disilane (Si2H6). For example, magnesia-cerammonium (Cp2Mg) can be selected as the doping precursor for the epitaxial growth of the AlN first nucleation layer 311.

[0252] Furthermore, the doping concentration of P-type impurities can be controlled by adjusting the carrier flow rate of the doping precursor. For example, the doping concentration of P-type impurities can be controlled by adjusting the carrier gas NH3.

[0253] In some feasible process flows, Mg can also be selected as a P-type impurity, and the concentration ρ of Mg can be 1×10⁻⁶. 17 cm -3 ≤ρ≤1×10 23 cm -3 For example, the concentration ρ of Mg can be 1×10⁻⁶. 22 cm -3 .

[0254] Furthermore, the growth temperature of the p-type doped first nucleation layer 311 is between 600°C and 1100°C, and may include 600°C and 1100°C, for example, approximately 1000°C.

[0255] like Figure 22 Step S11 in the middle, and Figure 23b , Figure 23b Is execution Figure 22 The structural diagram after step S12.

[0256] like Figure 22 Step S12 and Figure 23b A second nucleation layer 312 is formed on the first nucleation layer 311, the second nucleation layer 312 comprising a group IIIA nitride doped with a p-type impurity.

[0257] The P-type impurities here may all contain elements from Group II and / or at least one element from Group IVA.

[0258] Similar to the formation of the first nucleation layer 311 described above, epitaxial growth techniques can be used to prepare the second nucleation layer 312. For example, epitaxial growth techniques include, but are not limited to, metal-organic chemical vapor phase epitaxy (MOVPE), molecular beam epitaxy (MBE), or other single-crystal epitaxial growth techniques.

[0259] Furthermore, the preparation of the first nucleation layer 311 and the second nucleation layer 312 can be completed in the same device.

[0260] exist Figure 22 In step S12, the corresponding P-type impurity doped precursor may include, but is not limited to, diethylberyllium (DEBe), magnesium pyrocene (Cp2Mg), dimethylzinc (DMZn), calcium diisopropyl dithiocarbamate Ca(DPTC)2, methane (CH4), ethylene (C2H4), silane (SiH4), or silane (Si2H6), etc.

[0261] Figure 22 In step S12, the P-type impurity doped precursor can be combined with... Figure 22 In step S11, the P-type impurity doping precursor can be the same or different.

[0262] In some feasible process flows, mixed doping of C and Si can be selected, with C and Si concentrations ρ being 1 × 10⁻⁶. 15 cm -3 ≤ρ≤1×10 18 cm -3 For example, the concentrations ρ of C and Si can be 1×10⁻⁶. 17 cm -3 .

[0263] Furthermore, the growth temperature of the p-type doped first nucleation layer 311 is between 600°C and 1100°C, and may include 600°C and 1100°C, for example, approximately 1100°C.

[0264] like Figure 22 Step S13 in the middle, and Figure 23c , Figure 23c Is execution Figure 22 The structural diagram after step S13.

[0265] like Figure 22 Step S13 and Figure 23c A channel layer 32 is formed on the second nucleation layer 312.

[0266] The channel layer 32 can also be grown using epitaxial growth techniques, such as the aforementioned metal-organic chemical vapor phase epitaxy (MOVPE) and molecular beam epitaxy (MBE).

[0267] like Figure 22 Step S14 in the middle, and Figure 23d , Figure 23d Is execution Figure 22 The structural diagram after step S14 in the process.

[0268] like Figure 22 Step S14 and Figure 23d A barrier layer 33 is formed on the channel layer 32.

[0269] In some process flows, a stress buffer layer can be formed on the nucleation layer 31, which includes the first nucleation layer 311 and the second nucleation layer 312, before the channel layer 32 is formed, and then the channel layer 32 is formed on the stress buffer layer.

[0270] In some other processes, an isolation layer can be formed on the channel layer 32 before the barrier layer 33 is formed, and then the barrier layer 33 is formed on the isolation layer.

[0271] In some process flows, a cap layer can be formed on top of the barrier layer 33 after the barrier layer 33 is formed. This allows for the formation of a cap layer. Figure 23e The source (Source) 41, gate (Gate) 42, and drain (Drain) 43 are shown.

[0272] Using the above Figure 22 , Figures 23a to 23e In the fabricated radio frequency semiconductor device, the nucleation layer 31 includes a first nucleation layer 311 and a second nucleation layer 312 stacked in phase, and P-type impurities are doped in both the first nucleation layer 311 and the second nucleation layer 312 through a doping process, thereby suppressing the formation of N-type parasitic channels on the surface of the substrate 1.

[0273] This application also provides another method for fabricating a radio frequency semiconductor device, such as... Figure 24 ,and Figures 25a to 25e As shown. Figure 24 This is a process flow diagram of the fabrication method for this radio frequency semiconductor device. Figures 25a to 25e This is a cross-sectional view showing the completion of each step in the manufacturing process of this radio frequency semiconductor device.

[0274] like Figure 24 Step S21 in the middle, and Figure 25a , Figure 25a Is execution Figure 24 The structural diagram after step S21 in the process.

[0275] like Figure 24 Step S21 and Figure 25a A diffusion barrier layer 2 is formed on a substrate 1 containing silicon.

[0276] When forming the diffusion barrier layer 2, ammonia gas can be pre-passed through for nitriding pretreatment to form a silicon nitride thin film as the diffusion barrier layer 2 on the silicon-containing substrate 1; as another example, an alumina thin film can be formed on the silicon-containing substrate 1 as the diffusion barrier layer 2 using epitaxial growth technology.

[0277] like Figure 24 Step S22 and Figure 25b A first nucleation layer 311 is formed on the diffusion barrier layer 2. The first nucleation layer 311 includes a group IIIA nitride doped with a p-type impurity.

[0278] like Figure 24 Step S23 and Figure 25c A second nucleation layer 312 is formed on the first nucleation layer 311, the second nucleation layer 312 comprising a group IIIA nitride doped with a p-type impurity.

[0279] The P-type impurities here may all contain elements from Group II and / or at least one element from Group IVA.

[0280] like Figure 24 Step S24 and Figure 25d A channel layer 32 is formed on the second nucleation layer 312.

[0281] like Figure 24 Step S25 and Figure 25e A barrier layer 33 is formed on the channel layer 32.

[0282] like Figure 25f Source 41, Gate 42 and Drain 43 are located on barrier layer 33.

[0283] against Figure 24 The preparation methods of the first nucleation layer 311, the second nucleation layer 312, the channel layer 32 and the barrier layer 33 can refer to the preparation methods of the corresponding film structures in Figure 25 above.

[0284] Using the above Figure 24 , Figures 25a to 25f In the fabricated radio frequency semiconductor device, the nucleation layer 31 includes not only a first nucleation layer 311 and a second nucleation layer 312 stacked in phase, but also P-type impurities doped in both the first nucleation layer 311 and the second nucleation layer 312 through a doping process. Furthermore, a diffusion barrier layer 2 is formed between the substrate 1 and the nucleation layer 31. Therefore, not only can the formation of N-type parasitic channels on the surface of the substrate 1 be suppressed, but the formation of P-type parasitic channels on the surface of the substrate 1 can also be suppressed, thereby further reducing the radio frequency loss of the radio frequency semiconductor device.

[0285] In the description of this specification, specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0286] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A radio frequency semiconductor device, characterized in that, include: Substrate, comprising silicon; A nucleation layer is formed on the substrate; A channel layer is formed on the nucleation layer; A barrier layer is formed on the channel layer; The nucleation layer, the channel layer, and the barrier layer all contain group IIIA nitrides; The nucleation layer includes a first nucleation layer and a second nucleation layer stacked in a first direction, wherein the first nucleation layer is formed on the substrate and the second nucleation layer is formed on the first nucleation layer; Both the first nucleation layer and the second nucleation layer contain P-type impurities, and the P-type impurities contain at least one element from group II and / or group IVA; The concentration of the P-type impurity in the first nucleation layer is greater than the concentration of the P-type impurity in the second nucleation layer; The concentration of the P-type impurity in the first nucleation layer is ρ1; The concentration of the P-type impurity in the second nucleation layer is ρ2; Wherein, ρ1 / ρ2>10; The P-type impurities in the first nucleation layer are different from those in the second nucleation layer.

2. The radio frequency semiconductor device according to claim 1, characterized in that, Along the first direction, the thickness of the first nucleation layer is less than the thickness of the second nucleation layer.

3. The radio frequency semiconductor device according to claim 2, characterized in that, The thickness d1 of the first nucleation layer is: 1nm≤d1≤100nm; The thickness d2 of the second nucleation layer is: 100nm≤d2≤250nm.

4. The radio frequency semiconductor device according to claim 1, characterized in that, The dislocation density of the second nucleation layer is less than that of the first nucleation layer.

5. The radio frequency semiconductor device according to claim 1, characterized in that, The P-type impurities in the first nucleation layer and the P-type impurities in the second nucleation layer each contain at least two elements.

6. The radio frequency semiconductor device according to claim 1, characterized in that, The half-width at half-maximum (FWHM) of the rocking curves of the first nucleation layer and the second nucleation layer is less than or equal to 800 arcseconds.

7. The radio frequency semiconductor device according to claim 1, characterized in that, The ρ1 is: 1 10 17 cm -3 ≤ρ1≤1 10 20 cm -3 ; The ρ2 is: 1 10 15 cm -3 ≤ρ2≤1 10 18 cm -3 .

8. The radio frequency semiconductor device according to claim 1, characterized in that, The P-type impurities include at least one of beryllium, magnesium, calcium, zinc, carbon, and silicon.

9. The radio frequency semiconductor device according to claim 8, characterized in that, The P-type impurities in the first nucleation layer include magnesium; The P-type impurities in the second nucleation layer include at least two of magnesium, carbon, and silicon.

10. The radio frequency semiconductor device according to claim 1, characterized in that, The Group IIIA nitride in the first nucleation layer is the same as the Group IIIA nitride in the second nucleation layer.

11. The radio frequency semiconductor device according to any one of claims 1-10, characterized in that, The radio frequency semiconductor device further includes: a diffusion barrier layer; Along the first direction, the diffusion barrier layer is formed between the substrate and the first nucleation layer.

12. A method for fabricating a radio frequency semiconductor device, characterized in that, The preparation method includes: A first nucleation layer is formed on a silicon-containing substrate, the first nucleation layer comprising a group IIIA nitride doped with a p-type impurity; A second nucleation layer is formed on the first nucleation layer, the second nucleation layer comprising a Group IIIA nitride doped with a P-type impurity; The P-type impurity contains at least one element from group II and / or group IVA; The concentration of the P-type impurity in the first nucleation layer is greater than the concentration of the P-type impurity in the second nucleation layer; the concentration of the P-type impurity in the first nucleation layer is ρ1; the concentration of the P-type impurity in the second nucleation layer is ρ2; wherein, ρ1 / ρ2 > 10; The P-type impurities in the first nucleation layer are different from those in the second nucleation layer.

13. The method for fabricating a radio frequency semiconductor device according to claim 12, characterized in that, The precursors forming the first nucleation layer and / or the second nucleation layer include at least one of diethylberyllium, magnesium pyrocene, dimethylzinc, calcium diisopropyl dithiocarbamate, methane, ethylene, silane, and disilane.

14. The method for fabricating a radio frequency semiconductor device according to claim 12 or 13, characterized in that, Before forming the first nucleation layer on the silicon-containing substrate, the fabrication method further includes: A diffusion barrier layer is formed on a silicon-containing substrate to form the first nucleation layer on the diffusion barrier layer.

15. An electronic device, characterized in that, include: Circuit board; The radio frequency semiconductor device as described in any one of claims 1-11 or the radio frequency semiconductor device prepared by the method described in any one of claims 12-14; The radio frequency semiconductor device is disposed on the circuit board.

Images