High electron mobility transistor with improved performance and reliability and power amplifier including same

GaN-based HEMT devices with optimized geometries and two-stage biasing improve power amplifiers' efficiency and reliability, addressing the limitations of conventional materials and HEMTs for high power and frequency applications.

JP7591523B2Active Publication Date: 2024-11-28WOLFSPEED INC
View PDF 32 Cites 0 Cited by

Patent Information

Application Number
JP2021573429
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Priority Date
2019-08-29
Filing Date
2020-06-12
Publication Date
2024-11-28
Estimated Expiration
2040-06-12

AI Technical Summary

Technical Problem

Conventional semiconductor materials like silicon and gallium arsenide are expensive and not well-suited for high power and high frequency applications due to their small bandgaps and breakdown voltages, while existing high electron mobility transistors (HEMTs) face trade-offs between output power and bandwidth, with parasitic capacitance affecting switching speed.

Method used

The development of GaN-based HEMT devices with optimized geometries and configurations, including reduced via sizes and shared source contacts, along with a two-stage biasing system, to enhance power amplifiers' efficiency and reliability.

Benefits of technology

The GaN-based HEMT devices achieve high power density, improved power added efficiency, and reduced parasitic capacitance, enabling operation at frequencies up to 30 GHz with low quiescent current and high reliability under high-temperature reverse bias.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007591523000003
    Figure 0007591523000003
  • Figure 0007591523000004
    Figure 0007591523000004
  • Figure 0007591523000005
    Figure 0007591523000005
Patent Text Reader

Abstract

The power amplifier includes a GaN-based high electron mobility transistor (HEMT) device, and the power amplifier's power added efficiency (PAE) exceeds 32% at P1DB during operation of the power amplifier between 26.5 GHz and 30.5 GHz.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] This application claims priority to U.S. Patent Application No. 16 / 555,036, filed August 29, 2019, and U.S. Patent Application No. 16 / 440,427, filed June 13, 2019, the entire contents of each of which are incorporated herein by reference.

[0002] TECHNICAL FIELD The inventive concepts described herein relate to semiconductor devices, and more particularly to high electron mobility transistors. [Background technology]

[0003] Electronic devices formed of lower bandgap semiconductor materials, such as silicon and gallium arsenide, have found wide application in lower power and (in the case of silicon) lower frequency applications. However, these semiconductor materials may be less suitable for higher power and / or higher frequency applications due to their relatively small bandgaps (e.g., 1.12 eV for silicon and 1.42 eV for gallium arsenide at room temperature) and / or relatively small breakdown voltages.

[0004] High power, high temperature, and / or high frequency applications often use devices formed from wide bandgap semiconductor materials, such as silicon carbide (2.996 eV bandgap for alpha silicon carbide at room temperature) and Group III nitrides (e.g., 3.36 eV bandgap for gallium nitride at room temperature). These materials typically have higher electric field breakdown strengths and higher electron saturation velocities compared to gallium arsenide and silicon.

[0005] A device of particular interest for high power and / or high frequency applications is the High Electron Mobility Transistor (HEMT). HEMT devices can offer operational advantages in several applications. During operation, a two-dimensional electron gas (2DEG) is formed in a HEMT device at the heterojunction of two semiconductor materials with different band gap energies, the smaller band gap material having a higher electron affinity. The 2DEG is an accumulation layer in the smaller band gap material and can contain an extremely high sheet electron concentration. Furthermore, electrons originating from the wider band gap semiconductor material migrate to the 2DEG layer, allowing for high electron mobility due to reduced ionized impurity scattering. This combination of high carrier concentration and high carrier mobility can give HEMTs an extremely large transconductance and can provide a strong performance advantage over metal oxide semiconductor field effect transistors (MOSFETs) for high frequency applications.

[0006] High electron mobility transistors fabricated in III-nitride based material systems have the potential to generate large amounts of radio frequency (RF) power due to a combination of material properties including the aforementioned high breakdown field, wide band gap, large conduction band offset, and / or high saturated electron drift velocity.

[0007] Thus, HEMT devices can be advantageously utilized in power amplifiers (PAs). Power amplifiers, including those utilizing HEMT devices, often suffer from a trade-off between output power and bandwidth. Internal inefficiencies such as parasitic capacitances can affect the switching speed of the device. Smaller transistor geometries can achieve higher operating frequencies, but the smaller dimensions can result in reduced current (and power) capabilities.

[0008] Improving the amplifier's operating characteristics can help to efficiently use the available bandwidth. For example, linearity of a power amplifier can be useful in maintaining the integrity of complex modulation formats that may be used to achieve high data rates. For example, some modulation schemes may depend on the receiver being able to distinguish between variations in the amplitude of a signal, and linearity in a power amplifier helps preserve the amplitude variations of the signal. If the transmitted signal is distorted, for example, by gain compression of the power amplifier at a particular output power, the receiver may have difficulty decoding information based on the amplitude portion of the modulation. [Prior art documents] [Patent documents]

[0009] [Patent Document 1] U.S. Reissue Patent No. 34,861 [Patent Document 2] U.S. Pat. No. 4,946,547 [Patent Document 3] U.S. Patent No. 5,200,022 [Patent Document 4] U.S. Patent No. 6,218,680 [Patent Document 5] U.S. Patent No. 5,210,051 [Patent Document 6] U.S. Pat. No. 5,393,993 [Patent Document 7] U.S. Pat. No. 5,523,589 [Patent Document 8] U.S. Pat. No. 5,192,987 [Patent Document 9] U.S. Pat. No. 5,296,395 [Patent Document 10] U.S. Patent No. 6,316,793 [Patent Document 11] U.S. Patent No. 6,548,333 [Patent Document 12] U.S. Pat. No. 7,544,963 [Patent Document 13] U.S. Patent No. 7,548,112 [Patent Document 14] U.S. Patent No. 7,592,211 [Patent Document 15] U.S. Patent No. 7,615,774 [Patent Document 16] U.S. Patent No. 7,709,269 [Patent Document 17] U.S. Patent No. 7,030,428 [Patent Document 18] U.S. Pat. No. 8,563,372 [Patent Document 19] U.S. Pat. No. 9,214,352 [Patent Document 20] U.S. Pat. No. 8,049,252 [Patent Document 21] U.S. Patent No. 7,045,404 [Patent Document 22] U.S. Patent No. 8,120,064 [Patent Document 23] U.S. Patent No. 7,291,529 [Patent Document 24] U.S. Pat. No. 7,932,111 [Patent Document 25] U.S. Patent No. 7,259,402 [Patent Document 26] U.S. Pat. No. 8,513,686 [Patent Document 27] U.S. Patent Application Serial No. 11 / 080,905 [Non-patent literature]

[0010] [Non-Patent Document 1] “An Optical 0.25 μm GaN HEMT Technology on 100 mm SiC for RF Discrete and Foundry MMIC Products” by S. Wood et al. (2013 CS MANTECH Technical Digest, pp. 127-130, May 2013)

Non-Patent Document 2

Non-Patent Document 3

Non-Patent Document 4

Non-Patent Document 5

[0011] The embodiments described herein provide improved GaN-based HEMT devices, and improved power amplifiers including such HEMT devices.

[0012] According to some embodiments of the present invention, a power amplifier includes a GaN-based high electron mobility transistor (HEMT) device, and the power added efficiency (PAE) of the power amplifier is greater than 32% at output power at 1 dB compression (P1dB) during operation of the power amplifier between 26.5 GHz and 30.5 GHz.

[0013] In some embodiments, the power amplifier is configured to generate a power density of greater than 3.5 W / mm while operating at 30 GHz.

[0014] In some embodiments, the P1dB value of the power amplifier is within 1 dB of the maximum measured power during operation of the power amplifier between 26 GHz and 30.5 GHz.

[0015] In some embodiments, the ratio of quiescent drain current to power output is less than or equal to about 12.5 mA / W.

[0016] In some embodiments, the GaN-based HEMT device further includes a channel layer and a barrier layer stacked sequentially on the substrate, a source contact including a first ohmic portion on the barrier layer, and a via extending through the channel layer, the barrier layer, and the substrate to the first ohmic portion of the source contact.

[0017] In some embodiments, the combined thickness of the substrate, channel layer, and barrier layer is between 40 μm and 80 μm.

[0018] In some embodiments, the power amplifier further includes a first stage configured to perform gain compression and a second stage configured to perform gain expansion.

[0019] In some embodiments, the first stage is biased in class A mode and the second stage is biased in class AB mode.

[0020] In some embodiments, the drive ratio between the first stage and the second stage is approximately 1:2.

[0021] According to some embodiments of the present invention, a power amplifier includes a GaN-based high-electron-mobility transistor (HEMT) device configured to operate for more than 1000 hours in high-temperature-reverse-bias (HTRB) operation at 84V drain bias.

[0022] In some embodiments, the power added efficiency (PAE) of the power amplifier exceeds 32% at output power at 1 dB compression (P1dB) during operation of the power amplifier between 26.5 GHz and 30.5 GHz.

[0023] In some embodiments, the P1dB value of the power amplifier is within 1 dB of the maximum measured power during operation of the power amplifier between 26 GHz and 30.5 GHz.

[0024] In some embodiments, the HTRB operates at a temperature of 150°C.

[0025] In some embodiments, after 1000 hours of HTRB operation, the gate leakage of the HEMT device is less than 1×10 -4 From 1×10 -3 It is between A.

[0026] In some embodiments, the ratio of quiescent drain current to power output is less than or equal to about 12.5 mA / W.

[0027] In some embodiments, a GaN-based HEMT device includes a channel layer and a barrier layer stacked sequentially on a substrate, a source contact including a first ohmic portion on the barrier layer, and a via extending through the channel layer, the barrier layer, and the substrate to the first ohmic portion of the source contact.

[0028] In some embodiments, the combined thickness of the substrate, channel layer, and barrier layer is between 40 μm and 80 μm.

[0029] In some embodiments, the power amplifier further includes a gate contact on the barrier layer, the gate contact comprising a first portion and a second portion on the first portion, the second portion extending beyond one or more sidewalls of the first portion.

[0030] In some embodiments, the second portion of the gate contact extends beyond a sidewall of the first portion of the gate contact towards the source contact a distance between 0.5 μm and 0.15 μm.

[0031] In some embodiments, the power amplifier further includes a drain contact on the barrier layer, and the second portion of the gate contact extends beyond a sidewall of the first portion of the gate contact toward the drain contact a distance between 0.5 μm and 0.15 μm.

[0032] In some embodiments, the first portion of the gate contact includes a lower surface adjacent the barrier layer, and a length of the lower surface of the first portion of the gate contact is between 120 nm and 180 nm.

[0033] In some embodiments, the power amplifier further includes a first insulating layer over the barrier layer, and a second portion of the gate contact is on the first insulating layer.

[0034] In some embodiments, the power amplifier further includes a drain contact on the barrier layer and in the first insulating layer, a second insulating layer on the first insulating layer and the drain contact, and a field plate on the second insulating layer, the field plate extending beyond a sidewall of the second portion of the gate contact towards the drain contact a distance between 0.3 μm and 0.5 μm.

[0035] In some embodiments, the power amplifier is configured to operate in the Ka band.

[0036] According to some embodiments of the present invention, a GaN-based high electron mobility transistor (HEMT) device is configured to operate for more than 1000 hours at high temperature reverse bias (HTRB) operation at 84V drain bias.

[0037] In some embodiments, the HTRB operates at a temperature of 150°C.

[0038] In some embodiments, after 1000 hours of HTRB operation, the gate leakage of the HEMT device is less than 1×10 -4 From 1×10 -3 It is between A.

[0039] In some embodiments, the GaN-based HEMT further includes a channel layer and a barrier layer stacked sequentially on the substrate, a source contact including a first ohmic portion on the barrier layer, and a via extending through the channel layer, the barrier layer, and the substrate to the first ohmic portion of the source contact.

[0040] In some embodiments, the combined thickness of the substrate, channel layer, and barrier layer is between 40 μm and 80 μm.

[0041] In some embodiments, the GaN-based HEMT further includes a gate contact on the barrier layer, the gate contact comprising a first portion and a second portion on the first portion, the second portion extending beyond one or more sidewalls of the first portion.

[0042] In some embodiments, the second portion of the gate contact extends beyond a sidewall of the first portion of the gate contact towards the source contact a distance between 0.5 μm and 0.15 μm.

[0043] In some embodiments, the GaN-based HEMT further includes a drain contact on the barrier layer, and the second portion of the gate contact extends beyond a sidewall of the first portion of the gate contact toward the drain contact a distance between 0.5 μm and 0.15 μm.

[0044] In some embodiments, the first portion of the gate contact includes a lower surface adjacent the barrier layer, and a length of the lower surface of the first portion of the gate contact is between 120 nm and 180 nm.

[0045] In some embodiments, the GaN-based HEMT further includes a first insulating layer over the barrier layer, and a second portion of the gate contact is on the first insulating layer.

[0046] In some embodiments, the GaN-based HEMT further includes a drain contact on the barrier layer and in the first insulating layer, a second insulating layer on the first insulating layer and the drain contact, and a field plate on the second insulating layer, the field plate extending beyond a sidewall of the second portion of the gate contact toward the drain contact a distance between 0.3 μm and 0.5 μm.

[0047] In some embodiments, the HEMT device is configured to operate in the Ka band.

[0048] According to some embodiments of the invention, a high electron mobility transistor (HEMT) includes a substrate including a first surface and a second surface on opposite sides of the substrate, a channel layer on the first surface of the substrate opposite the substrate, a barrier layer on the channel layer, a source contact having a first ohmic contact on an upper surface of the barrier layer, and a via extending from the second surface of the substrate to the first ohmic contact.

[0049] In some embodiments, the HEMT further includes a backmetal layer within the via.

[0050] In some embodiments, the backmetal layer directly contacts the first ohmic contact.

[0051] In some embodiments, the back metal layer directly contacts a bottom surface of the first ohmic contact, and a contact area between the back metal layer and the bottom surface of the first ohmic contact is greater than or equal to 50 percent of the area of ​​the bottom surface.

[0052] In some embodiments, the back metal layer extends onto the second surface of the substrate.

[0053] In some embodiments, the via extends through the channel layer and the barrier layer.

[0054] In some embodiments, the substrate has a thickness of 75 μm or less.

[0055] In some embodiments, the maximum cross-section of the via in a direction parallel to the first surface of the substrate is less than 16 μm×40 μm.

[0056] In some embodiments, the maximum cross-sectional area of ​​the via in a direction parallel to the first surface of the substrate is less than 1000 μm 2 is less than.

[0057] In some embodiments, the first ohmic contact is in a recess in an upper surface of the barrier layer.

[0058] In some embodiments, the HEMT further includes a drain contact comprising a second ohmic contact on the top surface of the barrier layer.

[0059] In some embodiments, the HEMT further includes a gate contact on an upper surface of the barrier layer and a first insulating layer on an upper surface of the barrier layer, a first portion of the gate contact on the first insulating layer.

[0060] In some embodiments, the HEMT further includes a second insulating layer over the first insulating layer and the gate contact.

[0061] In some embodiments, the HEMT further includes a field plate on the second insulating layer, at least a portion of the field plate being on a portion of the second insulating layer between the gate contact and the drain contact.

[0062] In some embodiments, the substrate comprises silicon carbide.

[0063] In some embodiments, the barrier layer comprises a first Group III nitride and the channel layer comprises a second Group III nitride, and a first energy of a conduction band edge of the channel layer is less than a second energy of a conduction band edge of the barrier layer at an interface between the channel layer and the barrier layer.

[0064] In some embodiments, the HEMT further includes a doped well in the barrier layer adjacent to the first ohmic contact.

[0065] According to a further embodiment of the invention, a high electron mobility transistor (HEMT) includes a substrate, a semiconductor structure on the substrate, a first via and a second via each extending into the substrate and the semiconductor structure, a first ohmic source contact on the first via and a second ohmic source contact on the second via, a drain contact on the semiconductor structure between the first ohmic source contact and the second ohmic source contact, and a gate contact between the first source ohmic contact and the drain contact.

[0066] In some embodiments, the first via is between the first ohmic source contact and the substrate.

[0067] In some embodiments, the distance between the first ohmic source contact and the second ohmic source contact is less than or equal to 60 μm.

[0068] In some embodiments, the HEMT further includes a backmetal layer within the first via.

[0069] In some embodiments, the backmetal layer directly contacts the first ohmic source contact.

[0070] In some embodiments, the back metal layer directly contacts a bottom surface of the first ohmic source contact, and a contact area between the back metal layer and the bottom surface of the first ohmic source contact is greater than or equal to 50 percent of the area of ​​the bottom surface.

[0071] In some embodiments, the backmetal layer extends onto a surface of the substrate opposite the substrate from the semiconductor structure.

[0072] In some embodiments, the semiconductor structure includes a channel layer and a barrier layer on the channel layer.

[0073] In some embodiments, the first via extends through the channel layer and the barrier layer.

[0074] In some embodiments, the substrate has a thickness of 75 μm or less.

[0075] In some embodiments, the first via has a maximum cross-section in a direction parallel to the top surface of the substrate that is less than 16 μm×40 μm.

[0076] In some embodiments, the maximum cross-sectional area of ​​the first via in a direction parallel to the top surface of the substrate is less than 1000 μm 2 is less than.

[0077] In some embodiments, the first ohmic source contact is in a recess in a top surface of the semiconductor structure.

[0078] In some embodiments, the HEMT further includes a first insulating layer over the semiconductor structure, and the first portion of the gate contact is on the first insulating layer.

[0079] In some embodiments, the HEMT further includes a second insulating layer over the first insulating layer and the gate contact.

[0080] In some embodiments, the HEMT further includes a field plate on the second insulating layer, at least a portion of the field plate being on a portion of the second insulating layer between the gate contact and the drain contact.

[0081] According to a further embodiment of the invention, a method of forming a high electron mobility transistor (HEMT) includes forming a semiconductor structure on a substrate, the substrate including opposing upper and lower surfaces; forming a source contact comprising a first ohmic contact on the upper surface of the semiconductor structure; and forming a via extending from the lower surface of the substrate to the first ohmic contact.

[0082] In some embodiments, the method further comprises thinning the substrate.

[0083] In some embodiments, after thinning the substrate, the thickness of the substrate is 75 μm or less.

[0084] In some embodiments, the step of thinning the substrate is performed after forming the first ohmic contact and before forming the via.

[0085] In some embodiments, the method further includes depositing a backmetal layer in the via.

[0086] In some embodiments, the backmetal layer directly contacts the first ohmic contact.

[0087] In some embodiments, the back metal layer directly contacts the underside of the first ohmic contact, and a contact area between the back metal layer and the underside of the first ohmic contact is greater than or equal to 50 percent of an area of ​​the underside of the first ohmic contact.

[0088] In some embodiments, the backmetal layer extends over the bottom surface of the substrate.

[0089] In some embodiments, at least a portion of the via is between the first ohmic contact and the bottom surface of the substrate.

[0090] In some embodiments, forming a semiconductor structure on a substrate includes forming a channel layer on the substrate and forming a barrier layer on the channel layer.

[0091] In some embodiments, the maximum cross-section of the via in a direction parallel to the bottom surface of the substrate is less than 1000 μm 2 is less than.

[0092] In some embodiments, the method further includes forming a first insulating layer over the semiconductor structure, and forming the source contact includes forming a first ohmic contact in the first insulating layer.

[0093] In some embodiments, the method further includes forming a drain contact in the first insulating layer and forming a gate contact in the first insulating layer between the drain contact and the source contact.

[0094] In some embodiments, the drain contact includes a second ohmic contact.

[0095] In some embodiments, the method further includes forming a second insulating layer over the first insulating layer and the gate contact.

[0096] In some embodiments, the method further includes forming a field plate on the second insulating layer, at least a portion of the field plate being on a portion of the second insulating layer between the gate contact and the drain contact.

[0097] In some embodiments, the method further includes forming a metal contact in the second insulating layer on at least one of the gate contact, the drain contact, and the source contact.

[0098] According to a further embodiment of the invention, a high electron mobility transistor (HEMT) includes a substrate, a semiconductor structure on the substrate, a first ohmic source contact and a second ohmic source contact on the semiconductor structure, and a drain contact on the semiconductor structure between the first ohmic source contact and the second ohmic source contact, wherein a distance between the first ohmic source contact and the second ohmic source contact is less than or equal to 60 μm.

[0099] In some embodiments, the HEMT further includes a via extending from the bottom surface of the substrate to the first ohmic source contact.

[0100] In some embodiments, the HEMT further includes a backmetal layer within the via.

[0101] In some embodiments, the backmetal layer directly contacts the first ohmic source contact.

[0102] In some embodiments, the back metal layer directly contacts a bottom surface of the first ohmic source contact, and a contact area between the back metal layer and the bottom surface of the first ohmic source contact is greater than or equal to 50 percent of the area of ​​the bottom surface.

[0103] In some embodiments, the maximum cross-section of the via in a direction parallel to the bottom surface of the substrate is less than 16 μm×40 μm.

[0104] In some embodiments, the maximum cross-sectional area of ​​the via in a direction parallel to the bottom surface of the substrate is less than 1000 μm 2 is less than.

[0105] In some embodiments, the substrate has a thickness of 75 μm or less. [Brief description of the drawings]

[0106] [Figure 1A] FIG. 1 is a schematic plan view of a conventional high electron mobility transistor. [Figure 1B] FIG. 1B is a schematic cross-sectional view taken along line AA' in FIG. 1A. [Figure 1C] FIG. 1B is a schematic cross-sectional view taken along line BB' in FIG. 1A. [Figure 2A] 1 is a schematic plan view of a HEMT device according to an embodiment of the present invention. [Figure 2B] FIG. 2B is a schematic cross-sectional view of the HEMT device taken along line CC' in FIG. 2A. [Figure 2C] FIG. 2B is a schematic cross-sectional view of the HEMT device taken along line D-D' in FIG. 2A. [Diagram 3] 2B is a schematic cross-sectional view of an additional embodiment of a HEMT device taken along line D-D' of FIG. 2A, in accordance with an embodiment of the present invention. [Figure 4A] 1 illustrates a method of manufacturing a HEMT device according to an embodiment of the present invention. [Figure 4B] 1 illustrates a method of manufacturing a HEMT device according to an embodiment of the present invention. [Figure 4C] 1 illustrates a method of manufacturing a HEMT device according to an embodiment of the present invention. [Figure 4D] 1 illustrates a method of manufacturing a HEMT device according to an embodiment of the present invention. [Figure 4E] 1 illustrates a method of manufacturing a HEMT device according to an embodiment of the present invention. [Figure 4F] 1 illustrates a method of manufacturing a HEMT device according to an embodiment of the present invention. [Figure 4G] 1 illustrates a method of manufacturing a HEMT device according to an embodiment of the present invention. [Figure 4H] 1 illustrates a method of manufacturing a HEMT device according to an embodiment of the present invention. [Figure 4I] 1 illustrates a method of manufacturing a HEMT device according to an embodiment of the present invention. [Diagram 5] 2B is a schematic cross-sectional view of an additional embodiment of a HEMT device taken along line D-D' of FIG. 2A, in accordance with an embodiment of the present invention. [Figure 6A]FIG. 2 illustrates characteristics of an exemplary HEMT device according to embodiments described herein. [Figure 6B] FIG. 2 illustrates characteristics of an exemplary HEMT device according to embodiments described herein. [Figure 6C] FIG. 2 illustrates characteristics of an exemplary HEMT device according to embodiments described herein. [Figure 7] 4 is a graph of a gain curve of an example HEMT device according to embodiments described herein. [Figure 8] 1 is a graph of across-wafer small signal gain variation of an example HEMT device in accordance with embodiments described herein. [Figure 9A] 10 is a graph of a 10 GHz power sweep of an exemplary HEMT device according to embodiments described herein. [Figure 9B] 1 is a graph of a 30 GHz power sweep of an exemplary HEMT device according to embodiments described herein. [Figure 10A] FIG. 1 illustrates high temperature reverse bias (HTRB) test results of an exemplary HEMT device according to embodiments described herein. [Figure 10B] FIG. 1 illustrates high temperature reverse bias (HTRB) test results of an exemplary HEMT device according to embodiments described herein. [Figure 11A] FIG. 1 illustrates accelerated life test results for an exemplary HEMT device, according to embodiments described herein. [Figure 11B] FIG. 1 illustrates accelerated life test results for an exemplary HEMT device, according to embodiments described herein. [Figure 11C] FIG. 1 illustrates accelerated life test results for an exemplary HEMT device, according to embodiments described herein. [Figure 12] FIG. 2 illustrates the output power during accelerated life testing of an exemplary HEMT device, according to embodiments described herein. [Figure 13] FIG. 1 illustrates an example of a HEMT device that can provide performance improvements over conventional devices. [Figure 14] 1 is a graph of gain compression as a function of output power for a conventional GaN amplifier. [Figure 15] 1 is a schematic diagram of a multi-stage monolithic microwave integrated circuit (MMIC) device according to embodiments described herein. [Figure 16] 1 is a photograph of an exemplary MMIC according to embodiments described herein. [Figure 17A] 1 is a graph of S-parameters of a device according to embodiments described herein. [Figure 17B] 1 is a graph of S-parameters of a device according to embodiments described herein. [Figure 17C] 1 is a graph of S-parameters of a device according to embodiments described herein. [Figure 18A] 1 is a graph of power measurements for a device according to embodiments described herein. [Figure 18B] 1 is a graph of power measurements for a device according to embodiments described herein. [Figure 18C] 1 is a graph of power measurements for a device according to embodiments described herein. [Figure 19A] 1 is a graph of gain compression measurements of a device according to embodiments described herein. [Figure 19B] 1 is a graph of gain compression measurements of a device according to embodiments described herein. [Figure 19C] 1 is a graph of gain compression measurements of a device according to embodiments described herein. [Figure 20] 11 is a graph of maximum measured power and P1dB characteristics of a device according to embodiments described herein. [Figure 21] 1 is a graph of power added efficiency (PAE) characteristics of a device according to embodiments described herein. [Figure 22]1 is a graph of third-order intermodulation distortion (IM3) characteristics of a device according to embodiments described herein. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0107] Exemplary embodiments of the inventive concepts are described more fully below with reference to the accompanying drawings, in which embodiments of the invention are shown. The inventive concepts may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the inventive concepts to those skilled in the art. Like numbers refer to like elements throughout.

[0108] Embodiments of the present invention are directed to high electron mobility transistors that can exhibit improved device density and switching performance.

[0109] FIG. 1A is a schematic plan view of a conventional high electron mobility transistor 100, FIG. 1B is a schematic cross-sectional view of the high electron mobility transistor 100 taken along line A-A' in FIG. 1A, and FIG. 1C is a schematic cross-sectional view of the high electron mobility transistor 100 taken along line B-B' in FIG. 1A. As shown in FIGS. 1A-1C, the high electron mobility transistor 100 can be formed on a substrate 122, such as, for example, a silicon carbide substrate. A channel layer 124 is formed on the substrate 122. A barrier layer 126 is formed on the channel layer 124 opposite the substrate 122. The channel layer 124 can include, for example, gallium nitride (GaN), and the barrier layer 126 can include, for example, aluminum gallium nitride (AlGaN).

[0110] The channel layer 124 and the barrier layer 126 together may form a semiconductor structure 190 on the substrate 122. A source contact 215 and a drain contact 205 are formed on an upper surface of the barrier layer 126 and are laterally spaced apart from one another. The source contact 215 and the drain contact 205 may form ohmic contacts to the barrier layer 126.

[0111] A gate contact 210 is formed on top of the barrier layer 126 between the source contact 215 and the drain contact 205. A two-dimensional electron gas (2DEG) layer forms at the junction between the channel layer 124 and the barrier layer 126 when the HEMT device 100 is biased into its conducting or "on" state. The 2DEG layer acts as a highly conductive layer that allows current to flow between the source and drain regions of the device, which lie beneath the source and drain contacts 215 and 205, respectively.

[0112] The source contact 215 may be coupled to a reference signal, such as a ground voltage. The coupling to the reference signal may be achieved by a via 225 extending from the lower surface 122A of the substrate 122 through the substrate 122 to the upper surface 126A of the barrier layer. A metallization layer 217 may be deposited on the via 225 and on the two adjacent source contacts 215 to electrically connect the two adjacent source contacts 215. The metallization layer may be formed from a conductive metal.

[0113] The via 225 may expose a surface of the metallization layer 217. A backmetal layer 235 may be formed on the lower surface 122A of the substrate 122 and on the sidewalls of the via 225. The backmetal layer 235 may be in direct contact with the metallization layer 217. Thus, the backmetal layer 235 and the signal coupled thereto may be electrically connected to two adjacent source contacts 215.

[0114] The HEMT device 100 may be comprised of a plurality of transistor unit cells, examples of which are shown as 100_A and 100_B in FIG. 1A. Each of the transistor unit cells 100_A, 100_B may include a source contact 215, a drain contact 205, and a gate contact 210. A first ohmic source contact (e.g., 215_A) of a first one of the transistor unit cells 100_A may be separate from a second ohmic source contact (e.g., 215_B) of a second one of the transistor unit cells 100_B. The source contact 215, the drain contact 205, and the gate contact 210 may represent only a subset of the total number of source contacts 215, drain contacts 205, and gate contacts 210 of the HEMT device 100.

[0115] 1A, the size of the resulting HEMT device 100 may be affected by a source-to-source distance D1 between adjacent ones of the source contacts 215 and / or vias 225. As shown in FIGS. 1A and 1C, the distance D1 may be measured (e.g., in the X-direction of FIG. 1A) between adjacent ones of the vias 225 and / or between adjacent ones of a combination of two adjacent source contacts 215 interconnected by a metallization layer 217. The distance D1 may be a center-to-center distance in some embodiments. For example, in some embodiments, D1 may be measured from the center of a first via 225 to the center of an adjacent second via 225 in the X-direction. In conventional devices, the distance D1 may be 120 μm or more.

[0116] The present invention results from the realization that the dimensions of a semiconductor device can be reduced by directly connecting vias to source contacts without the need for an interconnect metallization layer, such as interconnect metallization layer 217 shown in Figures 1B and 1C. The substrate can also be thinned to allow for a reduction in the horizontal cross-sectional area of ​​the vias, further reducing the area of ​​the device. In addition to the improved density that can result from reducing the area of ​​the device, the overall parasitic capacitance can also be reduced. The reduced parasitic capacitance can lead to further improvements in the switching frequency of the resulting circuit.

[0117] FIG. 2A is a schematic plan view of a HEMT device 300, FIG. 2B is a schematic cross-sectional view of the HEMT device 300 taken along line C-C' in FIG. 2A, and FIG. 2C is a schematic cross-sectional view of the HEMT device 300 taken along line D-D' in FIG. 2A, in accordance with an embodiment of the present invention. FIGs. 2A, 2B, and 2C are intended to represent structures for identification and explanation purposes and are not intended to represent structures to physical scale. Furthermore, while FIGs. 2A, 2B, and 2C are shown to allow for general comparison with the structures of FIGs. 1A, 1B, and 1C, it will be understood that the scale of FIGs. 2A-2C is not intended to be the same as the scale of FIGs. 1A-1C.

[0118] As shown in Figures 2A-2C, a semiconductor structure 390, such as a III-Nitride HEMT semiconductor structure, can be formed on a substrate 322, such as a silicon carbide SiC substrate or a sapphire substrate. The substrate 322 can be a semi-insulating silicon carbide (SiC) substrate, which can be, for example, the 4H polytype of silicon carbide. Other possible silicon carbide polytypes can include the 3C, 6H, and 15R polytypes. The substrate can be a High Purity Semi-Insulating (HPSI) substrate available from Cree, Inc. The term "semi-insulating" is used descriptively herein, not in an absolute sense.

[0119] As used herein, the term "III-nitrides" refers to those semiconductor compounds formed between nitrogen (N) and elements in group III of the periodic table, usually aluminum (Al), gallium (Ga), and / or indium (In). The term also refers to ternary and quaternary (or higher) compounds, such as, for example, AlGaN and AlInGaN. As will be appreciated by those skilled in the art, group III elements can combine with nitrogen to form binary (e.g., GaN), ternary (e.g., AlGaN, AlInN), and quaternary (e.g., AlInGaN) compounds. All of these compounds have a formula in which one mole of nitrogen is combined with a total of one mole of group III elements.

[0120] In some embodiments of the present invention, the silicon carbide bulk crystal of the substrate 322 has a crystallinity of approximately 1×10 5 The SiC substrate may have a resistivity of ohm-cm or greater. Exemplary SiC substrates that may be used in some embodiments of the present invention may be manufactured, for example, by Cree, Inc. of Durham, North Carolina, the assignee of the present invention, and methods for manufacturing such substrates are described, for example, in U.S. Pat. Nos. 5,993,333, 5,993, 6,093, 6,093, 7,093, 7,093, 8,093, 9,093, 10,093, and 11,443, the disclosures of which are incorporated herein by reference in their entireties. Although silicon carbide may be used as the substrate material, embodiments of the present application may utilize any suitable substrate, such as sapphire (Al2O3), aluminum nitride (AlN), aluminum gallium nitride (AlGaN), gallium nitride (GaN), silicon (Si), GaAs, LGO, zinc oxide (ZnO), LAO, indium phosphide (InP), and the like. The substrate 322 may be a silicon carbide wafer, and the HEMT device 300 may be formed, at least in part, via wafer-level processing, and the wafer may then be diced to provide a plurality of individual high electron mobility transistors 300.

[0121] Substrate 322 may have a bottom surface 322A and a top surface 322B. In some embodiments, substrate 322 of HEMT device 300 may be a thin substrate 322. In some embodiments, the thickness of substrate 322 (e.g., in the vertical Z direction of FIG. 2B ) may be 100 μm or less. In some embodiments, the thickness of substrate 322 may be 75 μm or less. In some embodiments, the thickness of substrate 322 may be 50 μm or less.

[0122] A channel layer 324 is formed on an upper surface 322B of the substrate 322 (or on an optional layer described further herein), and a barrier layer 326 is formed on an upper surface of the channel layer 324. The channel layer 324 and the barrier layer 326 may each be formed by epitaxial growth in some embodiments. Techniques for epitaxial growth of III-nitrides are described, for example, in U.S. Pat. Nos. 5,399,413, 5,443, 5,513, 5,611, 5,714, 5,721, 5,811, 5,925, 5,933, 5,945, 5,951, 5,965, 5,103, 5,104, 5,105, 5,106, 5,107, 5,108, 5,109, 5,112, 5,114, 5,115, 5,109, 5,116, 5,117, 5,118, 5,122, 5,123, 5,132, 5,136, 5,141, 5,142, 5,143, 5,144, 5,145, 5,152, 5,154, 5,156, 5,158, 5,160, 5,162, 5,164, 5,166, 5,170, 5,172, 5,176, 5,178, 5,182, 5,190, 5,192, 5,193, 5,194, 5,195, 5,196, 5,197, 5,198,

[0123] In some embodiments, the channel layer 324 is Al, provided that the energy of the conduction band edge of the channel layer 324 is less than the energy of the conduction band edge of the barrier layer 326 at the interface between the channel layer 324 and the barrier layer 326. x Ga 1-x N (0≦x<1). In some embodiments of the invention, x=0, indicating that the channel layer 324 is GaN. The channel layer 324 may also be other III-nitrides, such as InGaN, AlInGaN, etc. The channel layer 324 may be undoped ("unintentionally doped") or may be grown to a thickness of greater than about 0.002 μm. The channel layer 324 may also be a multi-layer structure, such as a superlattice or a combination of GaN, AlGaN, etc. The channel layer 324 may be under compressive strain in some embodiments.

[0124] In some embodiments, the barrier layer 326 is AlN, AlInN, AlGaN, or AlInGaN, or a combination of layers thereof. The barrier layer 326 may include a single layer or may be a multi-layer structure. In certain embodiments of the invention, the barrier layer 326 may be thick enough and have a high enough aluminum (Al) composition and doping to induce a significant carrier concentration at the interface between the channel layer 324 and the barrier layer 326 by polarization effects when the barrier layer 326 is buried under an ohmic contact metal. The barrier layer 326 may be, for example, from about 0.1 nm to about 30 nm thick, but not so thick as to cause the formation of cracks or large defects therein. A barrier layer thickness in the range of 15-30 nm is typical. In some embodiments, the barrier layer 326 is undoped or doped with about 10 n-type dopants. 19 cm -3 In some embodiments of the present invention, the barrier layer 326 is doped to a concentration of less than Al. x Ga 1-xN is in the range of 0 < x < 1. In a specific embodiment, the aluminum concentration is about 25%. However, in other embodiments of the present invention, the barrier layer 326 includes AlGaN having an aluminum concentration between about 5% and less than about 100%. In a specific embodiment of the present invention, the aluminum concentration is greater than about 10%. The channel layer 324 and / or the barrier layer 326 can be deposited, for example, by metal-organic chemical vapor deposition (MOCVD), molecular beam epitaxy (MBE), or hydride vapor phase epitaxy (HVPE). As described above with respect to the conventional HEMT device 100, the 2DEG layer is induced in the channel layer 324 at the junction between the channel layer 324 and the barrier layer 326. The 2DEG layer functions as a highly conductive layer that enables conduction between the source region and the drain region of the device, which is under each of the source contact 315 and the drain contact 305. The channel layer 324 and the barrier layer 326 form a semiconductor structure 390.

[0125] Although the semiconductor structure 390 is shown with the channel layer 324 and the barrier layer 326 for illustrative purposes, the semiconductor structure 390 may include additional layers / structures / elements, such as a buffer layer and / or a nucleation layer between the channel layer 324 and the substrate 322, and / or a cap layer on the barrier layer 326. HEMT structures including substrates, channel layers, barrier layers, and other layers are illustratively described in U.S. Pat. Nos. 5,393,433, 5,393, 5,393, 6,496, 7,597, 7,598, 8,633, 8,751, 9,763, 9,825, 9,970, 9,982, 9,970 ... The optional buffer / nucleation / transition layers may be deposited by MOCVD, MBE, and / or HVPE.

[0126] The source contact 315 and the drain contact 305 may be formed on an upper surface 326A of the barrier layer 326 and may be laterally spaced apart from one another. The gate contact 310 may be formed on an upper surface 326A of the barrier layer 326 between the source contact 315 and the drain contact 305. The material of the gate contact 310 may be selected based on the composition of the barrier layer 326 and may be a Schottky contact in some embodiments. For example, nickel (Ni), platinum (Pt), nickel silicide (NiSi x Conventional materials capable of forming a Schottky contact to a gallium nitride based semiconductor material, such as GaN (GaN), Cu (Cu), palladium (Pd), chromium (Cr), tungsten (W), and / or tungsten silicon nitride (WSiN), may be used.

[0127] The source contact 315 and the drain contact 305 may include a metal capable of forming an ohmic contact to a gallium nitride based semiconductor material. Suitable metals include Ti, W, titanium tungsten (TiW), silicon (Si), titanium tungsten nitride (TiWN), tungsten silicide (WSi), rhenium (Re), niobium (Nb), Ni, gold (Au), aluminum (Al), tantalum (Ta), molybdenum (Mo), NiSi x , titanium silicide (TiSi), titanium nitride (TiN), WSiN, Pt, or other refractory metals. In some embodiments, source contact 315 may be an ohmic source contact 315. Thus, source contact 315 and drain contact 305 may include ohmic contact portions in direct contact with barrier layer 326. In some embodiments, source contact 315 and / or drain contact 305 may be formed from multiple layers to form ohmic contacts, which may be provided, for example, as described in commonly assigned U.S. Patent Nos. 5,993,333 and 5,993,621, the disclosures of which are incorporated herein by reference in their entireties.

[0128] The source contact 315 can be coupled to a reference signal, such as, for example, a ground voltage. The coupling to the reference signal can be achieved by a via 325 extending from the lower surface 322A of the substrate 322 through the substrate 322 to the upper surface 326A of the barrier layer. The via 325 can expose a bottom surface of the ohmic portion 315A of the source contact 315. A back metal layer 335 can be formed on the lower surface 322A of the substrate 322 and on the sidewalls of the via 325. The back metal layer 335 can be in direct contact with the ohmic portion 315A of the source contact 315. In some embodiments, the contact area between the back metal layer 335 and the bottom surface of the ohmic portion 315A of the source contact 315 can be 50 percent or more of the area of ​​the bottom surface of the ohmic portion 315A of the source contact 315. Thus, the back metal layer 335 and the signal coupled thereto can be electrically connected to the source contact 315.

[0129] In some embodiments, as shown in FIG. 2A, the via 325 can have an elliptical or circular cross-section when viewed in plan view, although the invention is not limited in this respect. In some embodiments, the cross-section of the via 325 can be polygonal or other shapes, as would be understood by one of ordinary skill in the art. In some embodiments, the dimensions (e.g., length and / or width) of the via are such that the maximum cross-sectional area A2 of the via 325 is less than 1000 μm 2 The cross-sectional area A2 may be taken in a direction parallel to the lower surface 322A of the substrate 322 (e.g., the XY plane in FIG. 2B). In some embodiments, the maximum cross-sectional area A2 of the via 325 may be a portion of the via 325 adjacent to the lower surface 322A of the substrate 322 (e.g., the opening of the via 325). For example, in some embodiments, the maximum width (e.g., in the X direction in FIG. 2A) may be 16 μm and the maximum length (e.g., in the Y direction in FIG. 2A) may be 40 μm, although the invention is not limited thereto. The maximum cross-sectional area A2 of the via 325 according to embodiments of the invention may be reduced from the maximum cross-sectional area of ​​a conventional device (such as that shown in FIGS. 1A-1C), in which the maximum cross-sectional area of ​​the via may be 35 μm by 75 μm or more. In some embodiments, the reduction in the cross-sectional area A2 of the via 325 may be achieved based at least in part on a reduced thickness of the substrate 322. In some embodiments, the sidewalls of the via 325 may be inclined and / or sloped relative to the bottom surface 322A of the substrate 322.

[0130] In some embodiments, the source contact 315, the drain contact 305, and the gate contact 310 may be formed as a plurality of source contacts 315, drain contacts 305, and gate contacts 310 on the substrate 322. Referring to FIG. 2A, the plurality of drain contacts 305 and source contacts 315 may be arranged alternately on the substrate 322. The gate contact 310 may be arranged between adjacent drain contacts 305 and source contacts 315 to form a plurality of transistor unit cells, examples of which are shown as 300_A and 300_B in FIG. 2A. Each of the transistor unit cells 300_A, 300_B may include a source contact 315, a drain contact 305, and a gate contact 310. An ohmic source contact (e.g., 315_AB) of a first one of the transistor unit cells 300_A may be shared with an ohmic source contact of a second one of the transistor unit cells 300_B. Although Fig. 2A shows a subset of the source contacts 315, drain contacts 305, and gate contacts 310 for ease of illustration, it will be understood that the HEMT device 300 may have additional structures including additional source contacts 315, drain contacts 305, and gate contacts 310 not shown in Fig. 2A.

[0131] In some embodiments, each of the plurality of source contacts 315, drain contacts 305, and gate contacts 310 on the substrate 322 may extend in a first direction (e.g., in the Y direction in FIG. 2A). A distance D2 measured (e.g., in the X direction in FIGS. 2A and 2C) between adjacent source contacts 315 having at least one drain contact 305 therebetween may be reduced from a conventional HEMT device, such as the HEMT device 100 shown in FIGS. 1A and 1B. In some embodiments, the distance D2 between adjacent source contacts 315 having a drain contact 305 therebetween may be 90 μm or less. In some embodiments, the distance D2 between adjacent source contacts 315 having a drain contact 305 therebetween may be 75 μm or less. In some embodiments, the distance D2 between adjacent source contacts 315 may be 60 μm or less. In some embodiments, the distance D2 between adjacent source contacts 315 may be 50 μm or less. In some embodiments, a plurality of vias 325 may be connected to each of the source contacts 315. The distance D2 may be a center-to-center distance between the source contacts 315 and / or between the vias 325 in some embodiments. For example, in some embodiments, the distance D2 may be measured from a center (e.g., in the X-direction) of a first source contact 315 to a center of an adjacent second source contact 315 in the X-direction. The distance D2 may extend from a center of the first source contact 315 across the drain contact 305 to a center of the second source contact 315. In some embodiments, the distance D2 may extend from a center of an ohmic portion of the first source contact 315 to a center of an ohmic portion of an adjacent second source contact 315 in the X-direction. In some embodiments, D2 may be measured from a center of a first via 325 connected to a first source contact 315 to a center of a second via 325 connected to an adjacent second source contact 315 in the X-direction.

[0132] As will be appreciated by those skilled in the art, a HEMT transistor may be formed by an active area between a source contact 315 and a drain contact 305 under control of a gate contact 310 between the source contact 315 and the drain contact 305. As shown in FIG. 2A, a HEMT device 300 may include adjacent HEMT unit transistors 300_A, 300_B that share a source contact 315. By sharing the source contact 315 between adjacent HEMT transistors and reducing the size of the via 325, a distance D2 measured between adjacent source contacts 315 with at least one drain contact 305 therebetween may be reduced from the source-to-source distance (see D1 in FIG. 1A and FIG. 1C) of a conventional HEMT device 100.

[0133] In some embodiments, the width of the source contact 315 and / or source region may also be reduced from a conventional HEMT device 100. In some embodiments, for example, the width of the source contact 315 (e.g., in the X-direction of FIGS. 2A-2C ) may be within 75% to 125% of the width of the drain contact 305. In other words, by reducing the size of the via 325, the width of the source contact 315 may be reduced between two adjacent gate contacts 310.

[0134] 2B and 2C, the HEMT device 300 may include a first insulating layer 350 and a second insulating layer 355. The first insulating layer 350 may be in direct contact with the top surface of the semiconductor structure 390 (e.g., in contact with the top surface 326A of the barrier layer 326). The second insulating layer 355 may be formed on the first insulating layer 350. It will also be appreciated that in some embodiments, three or more insulating layers may be included. The first insulating layer 350 and the second insulating layer 355 may function as passivation layers for the HEMT device 300.

[0135] The source contact 315, the drain contact 305, and the gate contact 310 may be formed in the first insulating layer 350. In some embodiments, at least a portion of the gate contact 310 may be on the first insulating layer. In some embodiments, the gate contact 310 may be formed as a T-gate and / or a gamma gate, the formation of which is illustratively described in U.S. Pat. Nos. 5,990,333; 5,990,353; and 6,133,627, the disclosures of which are incorporated by reference in their entireties herein. A second insulating layer 355 may be formed over the first insulating layer 350 and over portions of the drain contact 305, the gate contact 310, and the source contact 315.

[0136] A field plate 360 ​​may be formed on the second insulating layer 355. At least a portion of the field plate 360 ​​may be on the gate contact 310. At least a portion of the field plate 360 ​​may be on a portion of the second insulating layer 355 between the gate contact 310 and the drain contact 305. The field plate 360 ​​may reduce the peak electric field of the HEMT device 300, which may result in an increased breakdown voltage and reduced charge trapping. The reduced electric field may also result in other benefits, such as reduced leakage current and improved reliability. Field plates and techniques for forming field plates are illustratively described in U.S. Patent No. 6,399,992, the disclosure of which is incorporated herein by reference in its entirety.

[0137] Metal contacts 365 may be disposed within the second insulating layer 355. The metal contacts 365 may provide interconnections between the drain contact 305, the gate contact 310, and the source contact 315 and other portions of the HEMT device 300. Each of the metal contacts 365 may be in direct contact with each of the drain contacts 305 and / or the source contacts 315. The metal contacts 365 may include a metal or other highly conductive material, including, for example, copper, cobalt, gold, and / or composite metals. For ease of illustration, the second insulating layer 355, the field plate 360, and the metal contacts 365 are not shown in FIG. 2A.

[0138] 2A, 2B, and 2C result in grouping of HEMT transistors with reduced spacing. By sharing the source contacts 315 between adjacent ones of the HEMT transistors and reducing the size of the vias 325 connecting the source contacts 315 to the back metal layer 335, the spacing between adjacent source contacts 315 of the HEMT transistors can be reduced. The reduced spacing can result in a reduction in the overall parasitic capacitance of the HEMT device 300. The reduced parasitic capacitance can lead to further improvement in the switching frequency of the HEMT device 300.

[0139] In Figure 2C, for ease of illustration, the back metal layer 335 has been omitted and the configuration of the via 325 and semiconductor structure 390 is shown in dashed lines. Those skilled in the art will recognize that the reduced spacing D2 between adjacent source contacts may be achieved using other configurations of the via 325 and / or semiconductor structure 390 than those shown in Figures 2A-2C.

[0140] Figure 3 is a schematic cross-sectional view of an additional embodiment of a HEMT device 300' taken along line D-D' of Figure 2A, in accordance with an embodiment of the present invention. Portions of Figure 3 are substantially the same as or similar to those shown in Figures 2B and 2C, and duplicate descriptions thereof will be omitted for the sake of brevity. The HEMT device 300' may share a substantially similar plan view with the HEMT device 300 of Figure 2A, and therefore duplicate illustrations thereof will be omitted.

[0141] 3, the HEMT device 300' can have a source contact 315' recessed relative to a surface of the barrier layer 326. A bottom surface 315A' of the ohmic portion of the source contact 315' can be lower (e.g., closer to the substrate 322) than a top surface of the barrier layer 326. A sidewall and / or a side surface of the source contact 315' can be in contact with a portion of the barrier layer 326. In some embodiments, as shown in FIG. 3, a via 325 can extend through the barrier layer 326 and directly contact the bottom surface 315A' of the ohmic portion of the source contact 315'. A top portion of the back metal layer 335 in the via 325' (e.g., a portion of the back metal layer 335 furthest from the substrate 322) can be lower than a top surface 326A of the barrier layer 326. A top portion of the back metal layer 335 can be in direct contact with the bottom surface 315A' of the ohmic portion of the source contact 315'.

[0142] 3, only the source contact 315 is shown as recessed, however the invention is not so limited. In some embodiments, the gate contact 310 and / or the drain contact 305 may also be recessed relative to the top surface of the barrier layer 326.

[0143] 4A-4I illustrate a method of fabricating a HEMT device, such as HEMT device 300 of FIGS. 2A, 2B, and 2C, according to an embodiment of the invention.

[0144] 4A, a substrate 122 is provided upon which a semiconductor structure may be formed. A channel layer 324 is formed on the substrate 122, and a barrier layer 326 is formed on the channel layer 324. The substrate 122 may be a semi-insulating silicon carbide (SiC) substrate, which may be, for example, the 4H polytype of silicon carbide. Other silicon carbide polytype candidates may include the 3C, 6H, and 15R polytypes. The thickness of the substrate 122 may be 100 μm or greater.

[0145] Silicon carbide has a much closer crystal lattice match to III-nitrides (which may be used in the channel layer 324 and / or barrier layer 326) than sapphire (Al2O3), which may be a common substrate material for III-nitride devices. The closer lattice match may result in higher quality III-nitride films than those typically available on sapphire. Silicon carbide also has a relatively high thermal conductivity, and thus the total output power of III-nitride devices formed on silicon carbide may not be as limited by the heat dissipation of the substrate as similar devices formed on sapphire and / or silicon. Semi-insulating silicon carbide substrates may also provide device isolation and reduced parasitic capacitance.

[0146] Although silicon carbide may be used as the substrate, it should be understood that embodiments of the invention may utilize any suitable substrate for substrate 122, such as sapphire (Al2O3), aluminum nitride (AlN), aluminum gallium nitride (AlGaN), gallium nitride (GaN), silicon (Si), GaAs, LGO, zinc oxide (ZnO), LAO, indium phosphide (InP), etc.

[0147] Optional buffer, nucleation, and / or transition layers (not shown) may also be provided on the substrate 122. For example, an AlN buffer layer may be provided to provide a suitable crystal structure phase transition between the silicon carbide substrate and the remainder of the device. Additionally, a strain-balancing phase transition layer may also be provided.

[0148] 4A, a channel layer 324 is provided on the substrate 122. The channel layer 324 may be deposited on the substrate 122 using a buffer layer, a transition layer, and / or a nucleation layer as described above. The channel layer 324 may be under compressive strain. Furthermore, the channel layer 324 and / or the buffer layer, the nucleation layer, and / or the transition layer may be deposited by MOCVD, MBE, and / or HVPE. In some embodiments of the present invention, the channel layer 324 may be a III-nitride layer.

[0149] The barrier layer 326 may be a III-nitride layer. In some embodiments of the present invention, the barrier layer 326 may be a highly doped n-type layer. For example, the barrier layer 326 may be about 10 19 cm -3 It may be doped to a concentration of less than 1000 nm.

[0150] In some embodiments of the present invention, the barrier layer 326 may have a thickness, Al composition, and / or doping sufficient to induce a significant carrier concentration at the interface between the channel layer 324 and the barrier layer 326 by polarization effects when the barrier layer 326 is buried under an ohmic contact metal. The barrier layer 326 may also be thick enough to reduce or minimize scattering of electrons in the channel by ionized impurities deposited at the interface between the barrier layer 326 and a subsequently formed first protective layer.

[0151] In some embodiments, the channel layer 324 and the barrier layer 326 may have different lattice constants. For example, the barrier layer 326 may be a relatively thin layer having a smaller lattice constant than the channel layer 324, such that the barrier layer 326 "stretches" to the interface between the two layers. Thus, a pseudomorphic HEMT (pHEMT) device may be provided.

[0152] 4B, a first protective layer 410 is formed on the barrier layer 326. The first protective layer 410 is made of silicon nitride (Si x N yThe first protective layer 410 may be a dielectric material such as SiO2, aluminum nitride (AlN), silicon dioxide (SiO2), and / or other suitable protective materials. Other materials may also be utilized for the first protective layer 410. For example, the first protective layer 410 may also include magnesium oxide, scandium oxide, aluminum oxide, and / or aluminum oxynitride. Additionally, the first protective layer 410 may be a single layer or may include multiple layers of uniform and / or non-uniform composition.

[0153] The first protective layer 410 may be a blanket formed on the barrier layer 326. For example, the first protective layer 410 may be a silicon nitride (SiN) layer formed by high quality sputtering and / or plasma-enhanced chemical vapor deposition (PECVD). The first protective layer 410 may be thick enough to protect the underlying barrier layer 326 during subsequent annealing of the ohmic contacts.

[0154] Referring to FIG. 4C, the first protective layer 410 (see FIG. 4B) may be patterned to form the source contact 315, the gate contact 310, and the drain contact 305. For example, the first protective layer 410 may be patterned to form windows that expose the barrier layer 326 for placement of the source contact 315 and the drain contact 305. The windows may be etched using a patterned mask and a low damage etch to the barrier layer 326. An ohmic metal may be formed on the exposed portions of the barrier layer 326. The ohmic metal may be annealed to provide the source contact 315 and the drain contact 305.

[0155] The first protective layer 410 may also be patterned to provide the gate contact 310. The first protective layer 410 may be etched to form a window exposing the barrier layer 326 for placement of the gate contact 310. The gate contact 310 may be formed in the etched window and may extend through the first protective layer 410 to contact the exposed portion of the barrier layer 326. Suitable gate materials may depend on the composition of the barrier layer 326. However, in some embodiments, Ni, Pt, NiSi x Conventional materials capable of forming a Schottky contact to a nitride-based semiconductor material may be used, such as Cu, Pd, Cr, TaN, W, and / or WSiN. A portion of the gate contact 310 may extend onto a surface of the first protective layer 410. Formation of the source contact 315, gate contact 310, and drain contact 305 may result in patterning of the first protective layer 410 of FIG. 4B to form a first insulating layer 350.

[0156] 4C as being on top of the barrier layer 326, it will be understood that the source contact 315, the gate contact 310, and / or the drain contact 305 may be formed within a recess in the top surface of the barrier layer 326. Such a configuration may result in an embodiment similar to that shown in FIG.

[0157] 4D, a second protective layer 420 may be formed over the first insulating layer 350, the source contact 315, the gate contact 310, and the drain contact 305. The second protective layer 420 may be a dielectric layer. In some embodiments, the second protective layer 420 may have a different dielectric index than the first insulating layer 350.

[0158] 4E, a field plate 360 ​​may be formed on the second protective layer 420. The field plate 360 ​​may overlap the gate contact 310 (e.g., in the Z direction of FIG. 4E) and may extend a distance over the region between the gate and drain (i.e., the gate-drain region). The overlap of the field plate 360 ​​over the gate contact 310 and the distance that the field plate 360 ​​extends over the gate-drain region may be varied for optimal results. In some embodiments, the field plate 360 ​​may be electrically connected to the gate contact 310, and it is understood that field plate structures other than those shown in the figures may be used without departing from the invention.

[0159] Referring to FIG. 4F, the second protective layer 420 may be patterned to form metal contacts 365. For example, the second protective layer 420 may be patterned to form windows exposing the source contacts 315 and / or drain contacts 305 for placement of the metal contacts 365. The windows may be etched utilizing a patterned mask and a low damage etch to the source contacts 315 and / or drain contacts 305. A conductive metal may be formed on the exposed portions of the source contacts 315 and / or drain contacts 305 to form the metal contacts 365. Formation of the metal contacts 365 may result in patterning of the second protective layer 420 of FIG. 4E to form the second insulating layer 355.

[0160] Referring to FIG. 4G, the substrate 122 (see FIG. 4F) may be thinned to form a thinned substrate 322. In some embodiments, the thickness of the substrate 322 is reduced using a grinder, such as an in-feed or creep-feed grinder. In other embodiments, the thickness of the substrate 322 is reduced using lapping, chemical etching or reactive ion etching, with or without grinding, or a combination of these methods. In still other embodiments, etching may be used to treat the backside of the substrate 322 and reduce damage to the substrate 322 that may result from the thinning operation. Methods for thinning wafers are described, for example, in commonly assigned U.S. Patent Nos. 5,213,333; 5,213,333; and 5,213,333, the disclosures of which are incorporated herein by reference in their entireties.

[0161] In some embodiments in accordance with the invention, substrate 322 is thinned to a thickness between about 40 μm and about 100 μm, while in other embodiments, substrate 322 is thinned to a thickness between about 40 μm and about 75 μm.

[0162] 4H, a via 325 may be formed in the substrate 322, the channel layer 324, and the barrier layer 326. The via 325 may be formed by wet or dry etching. In some embodiments, the via 325 may be anisotropically etched such that a sidewall of the via 325 is sloped with respect to a top surface of the substrate 322. The via 325 may expose a bottom surface of the ohmic portion 315A of the source contact 315. In some embodiments, the source contact 315 may act as an etch stop material during the formation of the via 325.

[0163] With anisotropic etching, the maximum cross-sectional area A2 of the via 325 may be at a portion of the via 325 (e.g., at the opening of the via 325) adjacent the lower surface 322A of the substrate 322. In some embodiments, the anisotropic etching may result in the maximum cross-sectional area A2 of the via being related to the thickness of the substrate 322, since a thicker substrate 322 may result in a wider opening of the via 325. Thus, thinning the substrate 322, as described with respect to FIG. 4G, may provide additional benefit by reducing the cross-sectional area A2 of the via 325.

[0164] 4I, after formation of the vias, a back metal layer 335 may be deposited on the underside 322A of the substrate 322, the sidewalls of the vias 325, and the bottom surface of the ohmic portion 315A of the source contact 315. Thus, the back metal layer 335 may be in direct contact with the ohmic portion of the source contact 315. The back metal layer 335 may include a conductive metal such as, for example, titanium, platinum, and / or gold.

[0165] FIG. 5 is a schematic cross-sectional view of an additional embodiment of HEMT device 300″ taken along line D-D′ of FIG. 2A in accordance with an embodiment of the present invention. Portions of FIG. 5 are substantially the same as or similar to those shown in FIGS. 2B and 2C, and duplicate descriptions thereof will be omitted for the sake of brevity. HEMT device 300″ may share a substantially similar plan view with HEMT device 300 of FIG. 2A, and therefore duplicate illustrations thereof will be omitted.

[0166] 5, the HEMT device 300″ may have a well region 510 in the barrier layer 326 beneath one or more of the source contacts 315 and / or drain contacts 305. A bottom surface 315A of the ohmic portion of the source contact 315 and / or a bottom surface 305A of the ohmic portion of the drain contact 305 may be in contact with each of the well regions 510. The well region 510 may be formed by performing a doping implant on the barrier layer 326. In some embodiments, the well region 510 may be doped to have a higher doping concentration than the barrier layer 326. For example, when the barrier layer 326 is an n-type layer, the well region 510 may be doped with a higher concentration of n-type dopant (e.g., N + Or N ++ ) In some embodiments, the implantation of dopants to form well region 510 may be performed after formation of barrier layer 326 and before formation of source contact 315 and / or drain contact 305.

[0167] In some embodiments, the via 325 may be formed through the well region 510 beneath the source contact 315. Thus, a portion of the back metal layer 335 on the sidewall of the via 325 may directly contact a portion of one of the well regions 510. Although the well regions 510 are shown in FIG. 5 as being on both sides of the via 325, the invention is not so limited. In some embodiments, the well region 510 may be on only one side of the via 325. In some embodiments, the via 325 may be separated from the bottom surface 315A of the ohmic portion of the source contact 315, and a portion of the well region 510 may be disposed between the via 325 and the bottom surface 315A of the ohmic portion of the source contact 315.

[0168] 5 illustrates the via 325 as being in direct contact with the well region 510, the invention is not so limited. In some embodiments, isolation may be provided between the well region 510 and the via 325 and / or the back metal layer 335.

[0169] HEMT devices according to embodiments described herein may exhibit significant improvements over conventional devices. For example, HEMT devices according to embodiments described herein may exhibit output power density of greater than 3 W / mm and peak power added efficiency (PAE) of greater than 35% at 30 GHz. Small signal measurements at 28 V of devices according to embodiments described herein may exhibit a cutoff frequency of greater than 30 GHz and a 10 GHz small signal gain of greater than 20 dB. Furthermore, HEMT devices according to embodiments described herein may exhibit high voltage robustness under high temperature reverse bias (HTRB) of 84 V / 150° C., with 0 failures and negligible performance change after 1000 hours.

[0170] The HEMT device according to the embodiments described herein can include a substrate and epitaxy method that can be designed for RF performance at 30 GHz while maintaining robustness and high reliability. For example, the semiconductor structure of an exemplary HEMT device according to the embodiments described herein can include an AlN nucleation layer, an Fe-doped insulating GaN layer, an AlN barrier layer, and an undoped AlN barrier layer. 0.24 Ga 0.76 It may include a 15 nm cap layer of N.

[0171] 6A, 6B, and 6C show the characteristics of an exemplary HEMT device according to embodiments described herein. As shown in FIG. 6A, the source / drain ohmic contacts of the exemplary device are designed to provide a very smooth profile, which may enable consistent fine-line lithography and a well-controlled distribution of contact resistance centered around 0.2 Ω-mm. This is a reduction of 0.2 Ω-mm compared to a 0.25 μm process, which may result in an improvement in the peak current density, on-state resistance, and knee voltage of the exemplary device. FIG. 6B shows the gate length of each of the exemplary devices. The 150 nm gate length dimension (measured by atomic force microscope) shown in FIG. 6B was fabricated using an i-line lithography spacer method, such as that described in Non-Patent Document 1. The gate reduction technique resulted in gate lengths below the optical lithography limit while maintaining wafer-to-wafer repeatability. The ability to control the gate length using i-line lithography allows this process to be economically scalable beyond 100 mm diameter GaN on SiC substrates. FIG. 6C shows a cross section of a 150 nm gate of an exemplary device.

[0172] [Table 1]

[0173] Table 1 includes measurements taken from devices constructed according to the embodiments described herein. Typical electrical DC and RF characteristics shown in Table 1 were obtained on wafer. Small signal cutoff frequency and small signal gain were extracted from s-parameters at 2 GHz. A gain curve measured on a device with a gate circumference of 3.6 mm is shown in FIG. 7. The gain curve shows a maximum current density of over 1.2 A / mm and a peak transconductance of 470 mS / mm. The HEMT off-state characteristics are characterized by a pinch-off voltage of −2 V measured at 10 V and 1 mA / mm on the drain. The minimum breakdown voltage rating is 84 V at a gate voltage of −8 V and a maximum drain current of 1 mA / mm.

[0174] Multi-finger devices (6×75 μm) according to embodiments described herein were used to characterize the performance of 150 nm technology. These devices were biased at room temperature with a quiescent current and drain voltage of Id=100 mA / mm and Vd=28 V, respectively. FIG. 8 is a graph of the variation in Maximum Available Gain (MAG) of a sample device according to embodiments described herein. FIG. 8 shows that the across-wafer variation of small signal MAG measured on the 6×75 μm device was minimal across 20 test sites. As shown in FIG. 8, the across-wafer variation in small signal gain was a direct result of the ability to maintain a consistent gate length. FIGS. 9A and 9B are graphs of 10 GHz (FIG. 9A) and 30 GHz (FIG. 9B) power sweeps for 6×75 μm FETs biased at 28 V and 100 mA / mm. As shown in Figures 9A and 9B, continuous wave (CW) power sweeps were performed at 10 GHz and 30 GHz to determine the maximum output power (Pout) 910, gain 920, and peak power added efficiency (PAE) 930. The technique demonstrated power densities of >5 W / mm at 10 GHz and >3.5 W / mm at 30 GHz with PAE = 45% and 39%, respectively. Furthermore, when biased at 10 V, the device exhibits a cutoff frequency of 40 GHz.

[0175] The reliability and lifetime of the technology associated with the embodiments described herein was evaluated using 3.6 mm circumference devices to demonstrate the maturity of the technology. The HTRB system checks the ability of the device to withstand reverse bias while being subjected to increasing ambient temperatures that meet or exceed the maximum temperature that the part is rated to withstand. HTRB operation is described, for example, in U.S. Pat. No. 6,399,363, filed Mar. 15, 2005, entitled "GROUP III NITRIDE FIELD EFFECT TRANSISTORS (FETs) CAPABLE OF WITHSTANDING HIGH TEMPERATURE REVERSE BIAS TEST CONDITIONS," the disclosure of which is incorporated herein by reference as if fully set forth herein. Twenty devices were operated for 1000 hours on an HTRB system with a plate temperature of 150° C. Typical off-state bias conditions were Vg=−8V and Vd=84V. Figures 10A and 10B show the results of an 84V, 1000 hour HTRB test including in situ gate leakage (Figure 10A) and parameter variation of pinch-off voltage (Figure 10B). The in situ gate leakage was monitored throughout the test. The devices demonstrated negligible performance degradation and zero device failures, as shown in Figures 10A and 10B. For example, as shown in Figure 10A, at the completion of the HTRB test, the in situ gate leakage was 1 x 10 -4 From 1×10 -3 A. These examples provide validation of 150 nm Ka-band GaN HEMTs capable of withstanding 84 V HTRB for a standard 1000 hour test.

[0176] Preliminary reliability testing, including DC accelerated life test (DC-ALT), was performed for a junction temperature of 380° C. in large peripheral devices (3.6 mm). FIG. 11A and FIG. 11B show the parameter changes of pinch-off voltage (FIG. 11A) and on-resistance (FIG. 11B) during ALT at 380° C. FIG. 11C shows a 3.6 mm device according to an embodiment described herein after 400 hours of ALT testing. The DC-ALT bias conditions were Vd=28 V and 6 W / mm loss. FIG. 11A and FIG. 11B show the DC ALT reliability testing at 380° C. and the shift analysis of contact resistance and pinch-off voltage. All devices showed device stability throughout the testing. RF accelerated life testing (3.5 GHz) was performed at 325° C. in large peripheral devices (3.6 mm). 12 shows in situ monitoring of output power at 3.5 GHz RF ALT at 325° C. for an example device according to embodiments described herein. The monitored output power shown in FIG. 12 showed minimal degradation and zero failures after 400 hours.

[0177] The embodiments described herein include 150 nm GaN HEMT technology for foundry design of MMICs and can provide competitive performance up to 30 GHz. In some embodiments, the HEMT device can be biased at 28 V, although the embodiments described herein are not so limited. In some embodiments, the HEMT device can be biased at more than 15 V, for example 20 V. An optically defined i-line 150 nm gate can provide superior power and PAE for many Ka-band and Ku-band applications. The high voltage off-state HTRB reliability testing described herein confirms the robustness of the technology with negligible performance degradation. Multiple temperature accelerated life tests were used to evaluate the potential life of the technology.

[0178] FIG. 13 illustrates an exemplary embodiment of an improved HEMT device 600 according to embodiments described herein. Referring to FIG. 13, exemplary dimensions of device 600 are shown that provide the same and / or substantially similar performance as described herein. Device 600 may include elements described herein, particularly with respect to, for example, FIG. 2B, FIG. 2C, and FIG. 3. Accordingly, a duplicate description thereof will be omitted. In FIG. 13, elements sharing the same reference numbers as those of previous figures are intended to refer to the same or similar elements.

[0179] 13, a channel layer 324 may be formed on the substrate 322, and a barrier layer 326 may be formed on the channel layer 324. A first insulating layer 350 may be formed on the barrier layer 326, and a second insulating layer 355 may be formed on the first insulating layer 350. A thickness ID1 of the first insulating layer 350 may be between 100 nm and 150 nm. In some embodiments, the thickness ID1 of the first insulating layer 350 may be about 130 nm. In some embodiments, the first insulating layer 350 may include SiN. A thickness ID2 of the second insulating layer 355 may be between 180 nm and 220 nm. In some embodiments, the thickness ID2 of the second insulating layer 355 may be about 200 nm. In some embodiments, the first insulating layer 350 and the second insulating layer 355 may have a combined thickness of about 330 nm. The thickness of the first insulating layer 350 can affect the gate-source and gate-drain capacitances, which can affect the switching speed of the device 600. Similarly, the thickness of the second insulating layer 355 can affect the gate-drain capacitance, which can affect the switching speed and gain of the device 600.

[0180] The source contact 315 and the drain contact 305 may be formed in the first insulating layer 350 on the barrier layer 326. The source contact 315 may include an ohmic portion having a bottom surface 315A on the barrier layer 326. In some embodiments, the bottom surface 315A of the ohmic portion of the source contact 315 may extend below the top surface of the barrier layer 326, as shown in FIG. 3 . A via 325 may be formed in the substrate 322, the channel layer 324, and the barrier layer 326 to contact the bottom surface 315A of the ohmic portion of the source contact 315. The via 325 may include a back metal layer 335. In some embodiments, the thickness T of the wafer may be less than 100 nm. W (e.g., thicknesses of the substrate 322, channel layer 324, and barrier layer 326) may be between 40 μm and about 100 μm. In some embodiments, the wafer thickness T W In some embodiments, the wafer thickness T W The thickness of the wafer T may be about 75 μm. W The reduction in can enable the formation of vias 325 having smaller cross-sectional areas A2, which can reduce the overall size of device 600 and reduce inductance. In some embodiments, the maximum cross-sectional area A2 of vias 325 is less than 1000 μm 2 The following is the result.

[0181] The gate contact 310 may be formed in the first insulating layer 350, and a portion of the gate contact 310 may be under the second insulating layer 355. In Figure 13, the gate contact 310 is shown using a different shape than that of the gate contact 310 of Figures 2B, 2C, and 3, among others, for example. It will be understood that the present invention is not limited to a particular shape of the gate contact 310, and other shapes of the gate contact 310, as well as other elements of the device 600, are possible without departing from the embodiments described herein.

[0182] The gate contact 310 may include a first portion 310_1 in the first insulating layer 350 and a second portion 310_2 in the second insulating layer 355 over the first portion 310_1. A width of the second portion 310_2 of the gate contact 310 may exceed a width of the first portion 310_1. In some embodiments, the second portion 310_2 of the gate contact 310 may extend beyond one or more sidewalls of the first portion 310_1 of the gate contact 310 such that the gate contact 310 forms a T-gate and / or a gamma gate. On the drain side of the gate contact 310, the second portion 310_2 of the gate contact 310 extends a first distance Γ toward the drain contact 305 beyond the sidewalls of the first portion 310_1 of the gate contact 310. D The distance Γ D may be between 0.05 μm and 0.15 μm. In some embodiments, Γ D On the source side of the gate contact 310, the second portion 310_2 of the gate contact 310 is spaced a second distance Γ beyond the sidewall of the first portion 310_1 of the gate contact 310 toward the source contact 315. S The distance Γ S may be between 0.05 μm and 0.15 μm. In some embodiments, Γ S The first distance Γ may be about 0.1 μm. D may affect the gate-drain capacitance of the device 600, which may affect the switching speed and gain of the device 600. S may affect the gate-source capacitance of device 600, which may affect the switching speed of device 600.

[0183] The first portion 310_1 of the gate contact 310 may include a lower surface 310A adjacent to the barrier layer 326. G The length of the lower surface 310A of the gate contact 310, shown as L, may be between 120 nm and 180 nm.G The length L may be about 150 nm. G may represent the length of the interface between the gate contact 310 and the barrier layer 326. G may affect the transconductance and gate parasitic capacitance, which may affect the switching and efficiency of the device 600. The side and / or end of the lower surface 310A of the gate contact 310 closest to the drain contact 305 is spaced a distance L from the end (e.g., the closest end) of the drain contact 305, as shown in FIG. GD The distance L between the bottom surface 310A of the gate contact 310 and the drain contact 305 is GD In some embodiments, L may be between 1.75 μm and 2.25 μm. GD The side and / or edge of the underside 310A of the gate contact 310 closest to the source contact 315 may be a distance L from the edge (e.g., the closest edge) of the source contact 315, as shown in FIG. GS The distance L between the bottom surface 310A of the gate contact 310 and the source contact 315 can be GS In some embodiments, L may be between 0.5 μm and 0.7 μm. GS The distance L may be about 0.6 μm. GD and L GS may contribute to the on-resistance of device 600, which may affect device efficiency. In some embodiments, the gate contact 310 may be closer to the source contact 315 than to the drain contact 305.

[0184] A field plate 360 ​​may be formed on the second insulating layer 355 and the gate contact 310. A metal contact 365 may be formed in the second insulating layer 355 to electrically connect to the source contact 315 and the drain contact 305, respectively. On the drain side of the gate contact 310, the field plate 360 ​​extends a distance L 1 , as shown in FIG. 13 , beyond the sidewall of the second portion 310_2 of the gate contact 310 toward the drain contact 305. F2 In some embodiments, the distance L F2 may refer to the distance that the field plate 360 ​​extends beyond the outermost portion of the second portion 310_2 of the gate contact 310 that is closest to the drain contact 305. F2 In some embodiments, L may be between 0.4 μm and 0.5 μm. F2 The extension distance L of the field plate 360 ​​beyond the gate contact 310 may be about 0.4 μm. F2 may affect the gate-source capacitance and breakdown voltage of device 600, which may affect the voltage rating and switching speed of device 600.

[0185] One or more of the dimensions described with respect to FIG. 13 may contribute to improvements in device 600. In some embodiments, the configuration of device 600 may enable the device to achieve performance and improvements over conventional devices described herein. For example, embodiments described herein provide optically defined sub-200 nm (e.g., 120 nm to 200 nm) GaN HEMT technology with improved RF and reliability performance. For example, embodiments provided herein may have an output power density of 3.5 W / mm or greater at an operating frequency of 30 GHz. In some embodiments, the output power density at 30 GHz is between 3.5 W / mm and 4.0 W / mm. Embodiments provided herein may have a peak power added efficiency (PAE) of greater than 30% at an operating frequency of 30 GHz. Embodiments provided herein may have a transistor cutoff frequency of greater than 30 GHz and / or may be capable of performing 1000 hours HTRB at 150° C. with a drain bias of 84 V.

[0186] A 28V Ka-band MMIC was designed and fabricated using HEMT devices according to the embodiments described herein. The GaN MMIC has a bridge metal supported dielectric, 180 pF / mm 2The MMIC design utilizes microstrip lines on 75 μm (e.g., 3 mil) silicon carbide (SiC) with passive components such as 1000 Ω / sq. MIM capacitors, 12 Ω / sq. thin film resistors, and small through-wafer source slot vias. This MMIC design was used to demonstrate the ability of the technology described herein to meet the requirements of 5G communications mmWave and aerospace defense applications. The demand for greater bandwidth is creating more commercial applications for the Ka band. Satellite communications, point-to-point communications, 5G, and military radar are all applications that use or are planned to use the Ka band. The high power density and high breakdown voltage of GaN-on-SiC make it particularly interesting for power amplifiers. Most GaN power amplifiers begin to compress well before reaching saturation power. MMIC power amplifiers based on the embodiments described herein can exhibit high efficiency, low quiescent bias, and low gain compression, representing an improvement over the state of the art.

[0187] Conventional GaN amplifiers typically have a gradual gain compression that begins well before the saturation power is reached. A plot of gain compression versus output power (Pout) for a typical GaN amplifier is shown in FIG. 14. As shown in FIG. 14, the 1 dB compression point, also known as P1dB, is approximately 10 dB below the saturation power of the amplifier. P1dB, the output power at 1 dB compression, refers to the output power level at which the gate of the device is reduced by 1 dB from its constant (e.g., linear) value.

[0188] MMIC devices according to embodiments described herein may combine a gain compression driver stage with one or more gain expansion output stages. A schematic diagram of the MMIC device stages is shown in FIG. 15. Although FIG. 15 shows a three-stage amplifier having one gain compression stage and two gain expansion stages, embodiments described herein are not so limited. In some embodiments, the MMIC device may have more or less than three stages. In some embodiments, the gain compression portion of the MMIC device may include two or more stages. In some embodiments, the gain expansion portion of the MMIC device may include more or less than two stages.

[0189] In some embodiments, the gain compression driver stage may be biased in class A mode and / or the gain expansion stage may be biased in class AB mode. The gain compression stage may tend to offset the gain expansion of the output stage. By reducing the bias of the gain stage of a device close to pinch-off, the stage may have gain expansion and high efficiency. By increasing the bias of the gain stage, the stage may have gain compression and lower efficiency due to increased current. Thus, the output stage may be selected to be a low-biased gain expansion stage to help provide an improved combination of efficiency and low compression. These techniques, together with process improvements of the underlying device configuration, such as those described herein, result in an amplifier with significantly reduced gain compression, high efficiency, and low quiescent current. Examples of improved performance and related HEMT devices that can achieve this performance are described herein.

[0190] In some embodiments, MMIC devices according to embodiments described herein may provide improved RF performance. For example, in some embodiments, MMIC devices may provide output power at 30 GHz greater than 3.5 W / mm, peak power added efficiency (PAE) at 30 GHz greater than 30%, transistor cutoff frequency greater than 30 GHz, low gain compression, and / or power added efficiency at P1 dB greater than 30%. In some embodiments, MMIC devices according to embodiments described herein may provide improved reliability performance. For example, in some embodiments, MMIC devices may be capable of performing 1000 hours HTRB at 150° C. with a drain bias of 84 V.

[0191] MMIC devices according to embodiments described herein can be based on GaN-on-SiC devices. The GaN-on-SiC substrate and epitaxy method can be targeted for 30 GHz RF performance and 28V operation. The 150 nm gate can be optically defined using an i-line lithography spacer method. The 150 nm gate can be optimized by using a gate length (e.g., L shown in FIG. 13) on the device. G ), although the embodiments described herein are not limited to this gate length. In some embodiments, the gate length may be less than 200 nm (e.g., 120 nm to 180 nm). High temperature reverse bias (HTRB) stress testing was performed to verify the robustness and reliability of the process techniques described herein. HEMT devices were subjected to 84V HTRB for 1000 hours with negligible performance degradation and zero device failures. The thickness of the wafer (e.g., T W ) is 75 μm. Through-substrate vias, such as vias 325 described herein, are available and can be placed within 50 μm of each other. MMIC processes also have a 180 pF / mm 2 The MIM capacitors are 1000 Ω / sq, thin film resistors are 12 Ω / sq, and vias are in the HEMT source pad. DSS0 is approximately 700mA / mm.

[0192] The MMIC may be configured to operate at frequencies in the Ka band, which includes frequencies in the electromagnetic spectrum from 26.5 to 40 gigahertz (GHz). For example, the target frequency range for the MMIC PA may be 26.5 to 30.5 GHz, which may cover some potential 5G millimeter wave bands, although the embodiment is not so limited. In some embodiments, the MMIC device according to the embodiments described herein may operate in the Ku band (12 to 18 GHz). Other design goals for the MMIC may include a saturated power of greater than 4 W, a small signal gain of at least 25 dB, and a PAE of greater than 30% across the band. Additionally, low gain compression was desired, seeking a better P1dB than a typical three-stage GaN amplifier.

[0193] The MMIC PA may have three stages, as shown in Figure 15. In some embodiments, the first stage may be a 6x50 μm device, the second stage may have two 6x50 μm devices, and the third stage may have four 6x50 μm devices. This constitutes a 1:2 drive ratio for each stage.

[0194] A typical GaN amplifier has a large gain compression and can start compressing well before reaching the saturation power (Psat). The Psat value can provide a measurement of the amplifier's operating point where diminishing returns begin with respect to the gain provided. For example, above Psat, the gain provided for a given input may begin to decrease (i.e., exhibiting gain compression). Psat, as used herein, refers to the operating point on an amplifier where gain compression exceeds 3 dB. A design goal of some MMIC devices according to embodiments described herein was to improve 1 dB compressed power (P1dB) compared to a typical three-stage high efficiency GaN amplifier. To achieve this, the first stage may be biased separately from the last two stages and may be biased in class A mode, which may help increase the gain. The last two stages may be biased in class AB mode for efficiency. In some embodiments, the first stage is biased at 36 mA, which corresponds to 120 mA / mm. In some embodiments, the last two stages are biased at 20 mA / mm for a combined 36 mA. In some embodiments, the overall quiescent current of the amplifier is 72mA. Class A operation of the initial stages (e.g., the first stage in FIG. 15) may tend to result in gain compression, while class AB operation of the later stages (e.g., the last two stages in FIG. 15) may tend to result in gain expansion. These two different effects may be used to reduce the gradual gain compression typically found in GaN amplifiers. The drain voltage for design and operation may be 28V, although the embodiments described herein are not limited thereto.

[0195] The output load impedance can be determined by load pull data and circuit simulation. The matching network can have high pass and low pass elements to provide wideband operation. EM simulation and tuning can be used to improve the matching network for best overall circuit performance. FIG. 16 shows a photograph of a completed MMIC according to some embodiments described herein, however, this photograph is of an example device only and is not intended to limit the embodiments.

[0196] After wafer fabrication, S-parameters can be measured on the wafer. Figures 17A, 17B, and 17C are graphs of S-parameters for devices according to embodiments described herein. The S-parameters shown here are measured on 120 devices across the entire wafer. The lines on the graphs in Figures 17A-17C represent averages of the data. Figure 17A shows S21 across the wafer, Figure 17B shows S11 across the wafer, and Figure 17C shows S22 across the wafer. The S-parameters show very good broadband performance.

[0197] 18A, 18B, and 18C are graphs of power measurements of devices according to embodiments described herein. Power sweep tests were performed on five devices across a wafer from 26 GHz to 30.5 GHz. The power and PAE results presented are an average of these five devices. Power measurements were performed on the wafer using power pulses. The pulses had a duty cycle of 1% and a duration of 10 μs. These are typical on-wafer pulse output parameters to avoid overheating the wafer during testing. FIG. 18A shows the output power (Pout) versus input power (Pin) at 28.5 GHz in the center of the band. The associated gain versus input power is also plotted in FIG. 18A. The gain is relatively flat across the swept input power, which is desirable for linearity.

[0198] FIG. 18B shows the output power (Pout) at maximum PAE across an exemplary band of operating frequencies, with the corresponding maximum PAE shown in FIG. 18C. In the example shown in FIG. 18B, the targets of 4W output power and 30% PAE are exceeded. Furthermore, the example shown in FIG. 18C shows that the PAE is excellent across the band, approaching 40% at 30 GHz. It is highly desirable to compare the PAE measurements with previously published results. A comparison of power and PAE according to some embodiments described herein with previously published performance is shown in Table 2.

[0199] [Table 2] In Table 2, Reference A refers to Non-Patent Document 2, Reference B refers to Non-Patent Document 3, Reference C refers to Non-Patent Document 4, Reference D refers to Non-Patent Document 5, and Reference E refers to Non-Patent Document 6. As shown in Table 2, examples of the present disclosure can provide a quiescent drain current Idq of 72 mA for a maximum output power of 5.75 W, which can be converted to about 12.5 mA / W. In some embodiments, the ratio of the quiescent drain current to the power output is less than 12.5 mA / W.

[0200] 19A, 19B, and 19C are graphs of gain compression measurements of devices according to embodiments described herein. Gain compression measurements averaged over five devices are shown in FIG. 19A, 19B, and 19C at 26.5 GHz, 28.5 GHz, and 30.5 GHz, respectively. All figures show some gain expansion followed by gain compression. A typical GaN amplifier has a gradual gain compression that starts well before Psat. These plots show a rapid transition from P1dB to Psat. Gain expansion can affect linearity in the same way that gain compression can affect linearity. The amount of effect depends on the type of modulation. Parameters such as third order intermodulation distortion (IM3) are sensitive to the slope of gain versus input power. In the exemplary devices shown in FIG. 19A-19C, the observed gain expansion is moderate and has a gradual slope. As a result, the steep slope of gain compression still dominates.

[0201] FIG. 20 is a graph of maximum measured power (Pmax) and P1dB characteristics of a device according to an embodiment described herein. Maximum measured power (Pmax), as used herein, refers to the maximum power measured for the device. FIG. 20 compares P1dB and Pmax across frequency. As shown in FIG. 20, P1dB is within 1 dB of Pmax across the entire frequency band of a GaN MMIC PA according to an embodiment of the present invention. FIG. 21 is a graph of PAE characteristics of a device according to an embodiment described herein. The PAE with respect to Pmax and P1dB is shown in FIG. 21. For the exemplary device shown in FIG. 21, the PAE at P1dB is greater than 32% across the entire band, and the PAE at P1dB is 37.8% at 30 GHz. Over the operating band of 26.5 GHz to 30.5 GHz, the PAE at P1dB can range from 33% to about 40%. This represents an improved PAE at P1dB of a GaN-based power amplifier, such as a multi-stage MMIC according to an embodiment of the present invention, e.g., a three-stage GaN power amplifier MMIC operating at Ka-band, compared to a conventional device. The improved efficiency at P1dB of the present invention can be advantageously incorporated into a transmission system for communication signals utilizing digital modulation. Linearity is important in amplifiers utilized with modulated signals, and a higher P1dB with respect to output power and a higher efficiency (e.g., PAE) at P1dB can correlate to improved linearity at higher output power of the overall device.

[0202] FIG. 22 is a graph of IM3 characteristics of a device according to embodiments described herein. The IM3 of an example device according to embodiments described herein was measured under continuous waveform (CW) conditions, and the results are shown in FIG. 22. In a GaN amplifier, an IM3 of less than −25 dBc is advantageous. As shown in FIG. 22, this is achieved at a power level within 4 dB of the maximum power (Pmax) of the example device according to embodiments described herein, which is considered a good result.

[0203] Results have been described herein for an exemplary 28V, three-stage, Ka-band, GaN-on-SiC MMIC power amplifier using an optically defined i-line 150 nm gate. The MMIC according to the embodiments described herein can provide a maximum measured power (Pmax) of 37.6 dBm with an associated PAE of 39.8% at 30 GHz. The MMIC can exhibit low gain compression, excellent P1dB, and low quiescent bias. P1dB can be within 1 dB of Pmax across the 26.5-30.5 GHz band. At 30 GHz, P1dB can be 37.1 dBm with an associated PAE of 37.8% and a quiescent bias of 72 mA. This efficiency in P1dB represents an improvement over conventional devices for Ka-band multi-stage MMICs.

[0204] Terms such as first, second, etc. may be used herein to describe various elements, but it will be understood that these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first element can be referred to as a second element, and similarly, a second element can be referred to as a first element, without departing from the scope of the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0205] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the present invention. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises", "comprising", "includes" and / or "including" as used herein specify the presence of stated features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0206] Unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by those skilled in the art to which this invention belongs. It will be further understood that the terms used herein should be interpreted to have a meaning consistent with their meaning in the context of this specification and related art, and are not to be interpreted in an idealized or overly formal sense unless expressly so defined herein.

[0207] When an element, such as a layer, region, or substrate, is said to be "on" or extend "onto" another element, it will be understood that it can be directly on or extend directly onto the other element, or that intervening elements may also be present. In contrast, when an element is said to be "directly on" or extend "directly onto" another element, there are no intervening elements present. When an element is said to be "connected" or "coupled" to another element, it will be understood that it can be directly connected or coupled to the other element, or that intervening elements may be present. In contrast, when an element is said to be "directly connected" or "directly coupled" to another element, there are no intervening elements present.

[0208] Relative terms such as "lower" or "upper" or "top" or "lower" or "horizontal" or "lateral" or "vertical" may be used herein to describe the relationship of one element, layer, or region to another element, layer, or region as shown in the figures. It will be understood that these terms are intended to encompass various orientations of the device in addition to the orientation shown in the figures.

[0209] Embodiments of the present invention are described herein with reference to cross-sectional illustrations that are schematic illustrations of idealized embodiments (and intermediate structures) of the present invention. The thicknesses of layers and regions in the drawings may be exaggerated for clarity. Additionally, variations from the shapes of the illustrations are to be expected as a result, for example, of manufacturing techniques and / or tolerances. Thus, embodiments of the present invention should not be construed as limited to the particular shapes of regions illustrated herein but should include, for example, variations in shapes that result from manufacturing. Similarly, it will be understood that variations in dimensions are to be expected based on standard deviations in manufacturing procedures. As used herein, "about" includes values ​​within 10% of the nominal value.

[0210] Like numbers refer to like elements throughout. Thus, the same or similar numbers may be described with reference to other drawings even if they are not mentioned or described in the corresponding drawing. Additionally, elements not indicated by a reference number may be described with reference to other drawings.

[0211] Some embodiments of the present invention are described with reference to semiconductor layers and / or regions characterized as having a conductivity type, such as n-type or p-type, which refers to the majority carrier concentration within the layer and / or region. Thus, an N-type material has a majority equilibrium concentration of negatively charged electrons, while a P-type material has a majority equilibrium concentration of positively charged holes. Some materials may be designated with a "+" or "-" (as in N+, N-, P+, P-, N++, N--, P++, P--, etc.) to indicate a relatively larger ("+") or smaller ("-") concentration of majority carriers compared to another layer or region. However, such designations do not imply the presence of a particular concentration of majority or minority carriers within the layer or region.

[0212] In the drawings and specification, there are disclosed exemplary embodiments of the invention, and although specific terms have been employed, they are used in a generic and descriptive sense only and not for purposes of limitation, the scope of the invention being indicated in the following claims.

Claims

1. 1. A power amplifier comprising a GaN-based high electron mobility transistor (HEMT) device, the power amplifier having a power added efficiency (PAE) of greater than 32% at output power at 1 dB compression (P1 dB) during operation of the power amplifier between 26.5 GHz and 30.5 GHz; The GaN-based HEMT device comprises: a channel layer and a barrier layer sequentially deposited on a substrate to form active regions of a plurality of unit cell transistors of the GaN-based HEMT device; a source contact including a first ohmic portion on the barrier layer; a via extending from the substrate through the channel layer and the barrier layer to the first ohmic portion of the source contact, the via extending between active areas of two adjacent unit cell transistors of the GaN-based HEMT device; 2. A power amplifier comprising:

2. 10. The power amplifier of claim 1 configured to produce a power density of greater than 3.5 W / mm while operating at 30 GHz.

3. 3. A power amplifier as claimed in claim 1 or 2, wherein the P1dB value of the power amplifier is within 1 dB of maximum measured power during operation of the power amplifier between 26 GHz and 30.5 GHz.

4. a first stage configured to perform gain compression and a second stage configured to perform gain expansion; 4. The power amplifier of claim 1, wherein the first stage is biased in class A mode and the second stage is biased in class AB mode.

5. 5. A power amplifier as claimed in claim 1, having a ratio of quiescent drain current to power output of less than or equal to 12.5mA / W.

6. a gate contact on the barrier layer; the gate contact comprises a first portion and a second portion on the first portion; 6. A power amplifier as claimed in claim 1, wherein the second portion extends beyond one or more sidewalls of the first portion.

7. 7. The power amplifier of claim 6, wherein the second portion of the gate contact extends beyond a sidewall of the first portion of the gate contact toward the source contact a distance between 0.5 μm and 0.15 μm.

8. a drain contact on the barrier layer; 8. A power amplifier as claimed in claim 6 or 7, wherein the second portion of the gate contact extends beyond a sidewall of the first portion of the gate contact towards the drain contact a distance between 0.5 μm and 0.15 μm.

9. 2. The power amplifier of claim 1, wherein a cross-sectional area of ​​the via in a plane parallel to the substrate is smaller than a cross-sectional area of ​​the source contact in a plane parallel to the substrate.

Citation Information

Patent Citations

  • Cap layer including aluminum nitride for nitride-based transistor, and method of fabricating the same

    JP2008227501A

  • Method of forming vias in silicon carbide and resulting devices and circuits

    JP2008532290A

  • Passivation of wide band-gap based semiconductor devices with hydrogen-free sputtered nitrides

    JP2009055042A

  • Wide band power amplifier

    JP2012049909A

  • Group III nitride field effect transistors (FETS) capable of withstanding high temperature reverse bias test conditions

    US20060208280A1