Pseudomorphic high electron mobility transistor, low noise amplifier, and related devices
By adjusting the energy band structure of PHEMTs to keep the conduction band energy level below the Fermi energy at low output currents, the PHEMTs exhibit improved linearity and reduced intermodulation distortion, enhancing signal quality in wireless communication systems.
Patent Information
- Application Number
- JP2024535470
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2021-12-14
- Filing Date
- 2022-12-06
- Publication Date
- 2024-12-05
- Estimated Expiration
- 2042-12-06
AI Technical Summary
Existing pseudomorphic high electron mobility transistors (PHEMTs) suffer from intermodulation distortion due to nonlinearity, particularly at low output currents, which cannot be effectively suppressed by subsequent filters, leading to interference in wireless communication systems.
The PHEMT structure is modified by adjusting the energy band structure of the channel layer, including altering doping concentrations and thicknesses of layers, ensuring the conduction band energy level remains lower than the Fermi energy level at low output currents, thereby improving linearity and reducing intermodulation distortion.
This adjustment significantly enhances the linearity of PHEMTs, as evidenced by increased output 3rd-order intercept point (OIP3) values, thereby improving signal quality and reducing interference in low noise amplifiers.
Smart Images

Figure 2024545219000001_ABST
Abstract
Description
[Technical field]
[0001] pseudo child The application relates to the field of semiconductor device technology, and in particular to pseudomorphic high electron mobility transistors, low noise amplifiers, and related devices. [Background technology]
[0002] As shown in Figure 1, when the received signal passes through the low noise amplifier (LNA), the received signal on the receiving link at the antenna end of the wireless communication product and the transmitted signal leaking into the receiving link will generate an intermodulation signal, resulting in intermodulation distortion (IMD). This is caused by the nonlinear characteristics of the LNA. Because the frequency of the third-order intermodulation product is very close to the frequency of the received signal, the third-order intermodulation product cannot be suppressed by the subsequent filter in the receiving link, causing interference to the received signal on the link. Summary of the Invention
[0003] The embodiments of this application provide a pseudomorphic high electron mobility transistor (PHEMT), a low noise amplifier, and related devices. When the output current of the PHEMT is smaller than a first threshold, the conduction band energy level of the channel layer is lower than the Fermi energy level. This improves the linearity of the PHEMT under small output currents and reduces the intermodulation distortion caused by the LNA nonlinearity.
[0004] According to a first aspect, an embodiment of the present application provides a pseudomorphic high electron mobility transistor (PHEMT), the PHEMT comprising: A channel layer; a lower barrier layer and an upper barrier layer disposed on either side of the channel layer, the lower barrier layer being connected to the channel layer; a first isolation layer and a first doped layer disposed between the channel layer and the upper barrier layer, the first isolation layer configured to isolate the first doped layer from the channel layer, the first doped layer configured to provide a two-dimensional electron gas; Including, When the output current of the PHEMT is smaller than the first threshold, the conduction band energy level of the channel layer is lower than the Fermi energy level.
[0005] An energy level structure of a channel layer in the PHEMT is adjusted such that a conduction band energy level of the channel layer is lower than a Fermi energy level when an output current of the PHEMT is smaller than a first threshold, which improves the linearity of the PHEMT under small output currents and reduces intermodulation distortion caused by LNA nonlinearity.
[0006] In one possible implementation, the lower barrier layer is directly connected to the channel layer.
[0007] The removal of the lower doped layer in the PHEMT results in a smaller gradient of the energy level of the channel layer in the thickness direction, and the conduction band energy level of the channel layer is lower than the Fermi energy level during operation under small currents.
[0008] In one possible implementation, the first doped layer is silicon doped and has a doping concentration of 3e12 cm -2 From 5e12cm -2 It is.
[0009] In one possible implementation, the doping concentration of the first doped layer is instead , 5e12cm -2 From 6e12cm -2 Yes It is okay .
[0010] Increasing the concentration of the first doped layer in the PHEMT increases the gain of the PHEMT.
[0011] In one possible implementation, the lower barrier layer is connected to the channel layer via a second isolation layer and a second doped layer, the second isolation layer configured to isolate the channel layer from the second doped layer, and the second doped layer configured to provide a two-dimensional electron gas.
[0012] The PHEMT is a double-doped PHEMT, for example a double δ-doped PHEMT.
[0013] In one possible implementation, the doping concentration of the first doped layer is 3.5e12 cm -2 From 4.5e12cm -2 and the doping concentration of the second doped layer is 3e11cm -2 From 5e11cm -2 It is.
[0014] In one possible implementation, the doping concentration of the first doped layer is 4e12 cm -2 From 6e12cm -2 and the doping concentration of the second doped layer is 2e8cm -2 From 3e11cm -2 It is.
[0015] In one possible implementation, a ratio of the doping concentration of the first doped layer to the doping concentration of the second doped layer is greater than a preset value.
[0016] Optionally, the preset value is a positive number greater than 6, for example 9, 10, 15, 30, 70, 100, or 150.
[0017] If the total doping concentration of the PHEMT remains unchanged or is not decreased, the doping concentration of the lower doped layer (second doped layer) is decreased or the ratio of the doping concentration of the upper doped layer (first doped layer) to the doping concentration of the lower doped layer (second doped layer) is increased, so that the gradient of the energy level of the channel layer in the thickness direction becomes smaller and the conduction band energy level of the channel layer becomes lower than the Fermi energy level during operation under small current.
[0018] In one possible implementation, the thickness of the channel layer is 15-20 nm, 18-20 nm, or 20-25 nm, for example 18 nm.
[0019] In one possible implementation, when the output current of the PHEMT is less than a second threshold, the conduction band energy level of the channel layer is substantially decreased in the thickness direction.
[0020] The thickness of the channel layer of the PHEMT is increased. R, When the device works To Under the channel layer Electrons stored in is avoided By doing so, Electrons More uniform This improves the linearity of the PHEMT.
[0021] In one possible implementation, the PHEMT further includes a cap layer, a source, a drain, and a gate, the cap layer being disposed on a side of the upper barrier layer away from the channel layer and having a through hole for providing an ohmic contact, the gate being disposed within the through hole, and the source and drain being both disposed on a side of the cap layer away from the upper barrier layer and located on either side of the through hole.
[0022] In one possible implementation, the channel layer is made of indium gallium arsenide and the top barrier layer, bottom barrier layer, or separation layer is made of aluminum gallium arsenide.
[0023] According to a second aspect, an embodiment of the present application further provides a low noise amplifier including a PHEMT according to the first aspect or any one of the possible implementations of the first aspect.
[0024] According to a third aspect, an embodiment of the present application further provides a radio frequency circuit including a low noise amplifier according to the second aspect.
[0025] According to a fourth aspect, an embodiment of the present application further provides a radio frequency chip including at least one of a PHEMT according to the first aspect or any one of possible implementations of the first aspect, a low noise amplifier according to the second aspect, and a radio frequency circuit according to the third aspect.
[0026] According to a fifth aspect, an embodiment of the present application further provides an electronic device including at least one of a PHEMT according to the first aspect or any one of possible implementations of the first aspect, a low noise amplifier according to the second aspect, a radio frequency circuit according to the third aspect, and a radio frequency chip according to the fourth aspect. [Brief description of the drawings]
[0027] In order to more clearly describe the technical solutions in the embodiments of this application, the following briefly describes the accompanying drawings for illustrating the embodiments. [Figure 1] 2 is a schematic illustration of an intermodulation signal generated on an antenna-end receive link according to an embodiment of the present application; [Figure 2A] FIG. 1 is a schematic diagram of the structure of a double δ-doped PHEMT according to the prior art. [Figure 2B] FIG. 2B is a diagram showing an example of a DC output characteristic curve of the PHEMT shown in FIG. 2A. [Figure 2C] FIG. 2B is a schematic diagram of the energy band structure of the PHEMT shown in FIG. 2A. [Figure 2D] 2B is an example of a curve diagram of OIP3 value of the PHEMT shown in FIG. 2A under different output currents. [Diagram 3]FIG. 1 is a schematic diagram defining OIP3 values in the prior art; [Figure 4] FIG. 1 is a schematic diagram of a cross-sectional structure of a single δ-doped PHEMT according to an embodiment of the present application. [Diagram 5] FIG. 2 is a schematic diagram of a cross-sectional structure of a double δ-doped PHEMT according to one embodiment of the present application. [Figure 6A] 6A and 6B show example diagrams of OIP3 value curves of a double δ-doped PHEMT according to the prior art and a single δ-doped PHEMT according to an embodiment of this application under different output currents. [Figure 6B] 6A and 6B show example diagrams of OIP3 value curves of a double δ-doped PHEMT according to the prior art and a single δ-doped PHEMT according to an embodiment of this application under different output currents. [Figure 7A] 1 is an example of an energy band diagram of a double δ-doped PHEMT according to the prior art; [Figure 7B] FIG. 2 is an example of an energy band diagram of a single δ-doped PHEMT according to an embodiment of the present application. [Figure 7C] FIG. 2 is an example of a diagram of the conduction band structure of a single δ-doped PHEMT and a double δ-doped PHEMT in a channel layer according to an embodiment of the present application. [Figure 8] FIG. 13 shows the electron concentration distribution of double δ-doped PHEMTs with channel layers of 12 nm and 18 nm thickness, respectively, under an output current of 60 mA / mm. [Figure 9] 1 is a schematic diagram of a circuit structure of an LNA according to an embodiment of the present application. [Figure 10] FIG. 1 is a schematic diagram of a radio frequency system according to an embodiment of the present application. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0028] In general, there are two methods for reducing the interference caused to the received signal by intermodulation signals.
[0029] In one method, the isolation of the duplexer is increased to suppress the leakage of the transmit signal to the receive link. The duplexer, also called a diplexer, functions to isolate the transmit signal transmitted by the shared antenna from the receive signal received by the shared antenna to ensure that both the transmit link and the receive link can operate simultaneously. If the isolation of the duplexer is increased to suppress the interference caused by the intermodulation signal to the receive signal, the cavity of the duplexer needs to be made larger. This increases the size of the duplexer and also increases the cost significantly.
[0030] Another method is to improve the linearity of a low noise amplifier (LNA) to reduce intermodulation signals that are generated by the leakage signal and the received signal and cannot be completely suppressed by the filter. The linearity of the LNA is determined by the core of the transistor. The transistor may be a pseudomorphic high electron mobility transistor (PHEMT).
[0031] Figure 2A shows a typical structure of a gallium arsenide (GaAs)-based PHEMT. The epitaxial layers of the PHEMT core are, from top to bottom, the buffer layer, Lower barrier layer, lower δ-doped layer, lower isolation layer, channel layer,The upper isolation layer, the upper delta-doped layer, the upper barrier layer, and the cap layer. For a gallium arsenide (GaAs)-based PHEMT, the buffer layer on the substrate may be made of gallium arsenide or aluminum gallium arsenide and is configured to prevent defects in the substrate from entering other layers of the PHEMT. The lower barrier layer and the upper barrier layer may be made of aluminum gallium arsenide (AlGaAs) and are configured to prevent two-dimensional electron gas from entering the buffer layer. Both the lower delta-doped layer and the upper delta-doped layer may be silicon (Si) doped and are configured to provide a two-dimensional electron gas. The upper isolation layer and the lower isolation layer may be made of aluminum gallium arsenide (AlGaAs) and are configured to separate the doped layer from the channel layer. The channel layer may be made of indium gallium arsenide (InGaAs). The cap layer may be made of heavily doped gallium arsenide and connects the source and drain. The electron concentration in the channel layer is controlled by using the gate voltage so that the output current between the source and the drain can be controlled. This realizes the signal amplification function. Figure 2B shows the DC output characteristic curve of the PHEMT shown in Figure 2A. For example, five different gate voltages are shown in Figure 2B to distinguish five operating states of the PHEMT, namely, the off state (1), the on-state (2), the on state (3), the fully on state (4), and the linear operating state (5), respectively.
[0032] 2C is a schematic diagram of the energy band structure of the PHEMT shown in FIG. 2A in five different operating states. It should be understood that each energy band structure diagram in FIG. 2C simply shows the conduction band structure of all layers, including the conduction bands of the cap layer, upper barrier layer, upper δ-doped layer, upper isolation layer, channel layer, lower isolation layer, lower δ-doped layer, and lower barrier layer from left to right, i.e., in the thickness direction. Only the positions of the upper δ-doped layer, lower δ-doped layer, and channel layer are marked in the diagram.
[0033] As shown in FIG. 2C, when the PHEMT is in the off state (1), the conduction band energy level of the channel layer is E F As the gate voltage increases, when the PHEMT is in the on-state (2), the conduction band energy level of the channel layer rises to the Fermi energy level E F As the gate voltage increases further, the conduction band energy level of the channel layer partially approaches the Fermi energy level. E F When the gate voltage is lower than , a certain amount of free electrons appears in the channel layer. In this case, the PHEMT is in the on state (3). As the gate voltage is further increased, the conduction band structure of the channel layer changes from the left-sloping, right-sloping form in the on state (3) to the form in the fully on state (4) and then to the form in the linear operation state (5). That is, carriers tend to be located on the upper side of the channel layer (i.e., the left side in FIG. 2D, i.e., the side adjacent to the upper separation layer).
[0034] However, to meet the requirements of low power consumption and low noise in the receive link of a radio frequency system, the LNA is substantially operated with the output current density limited to 60-100 mA / mm, i.e., the on-state (3) of the PHEMT in FIG. 2C.
[0035] In order to meet the requirements of low power consumption and low noise factor of LNA in radio frequency circuits, the output current of the operating point is practically limited to 60-100mA / mm. Due to the characteristics of high electron mobility, PHEMT can reach the operating state current only with a low concentration of two-dimensional electron gas. As a result, most of the operating points are near the turn-on voltage of PHEMT. However, the linearity of PHEMT in this range is small and cannot meet the requirements of the application. The linearity of PHEMT is represented by the output 3rd order intercept point (OIP3). Higher OIP3 indicates higher linearity. As shown in FIG. 2D, the curve of OIP3 value of PHEMT shown in FIG. 2A under different output currents can be obtained through calculation based on the DC output characteristic curve shown in FIG. 2B. It can be seen from FIG. 2D that when the output current is 60mA, the OIP3 value of PHEMT is about 36dBm, which cannot meet the requirement of 41dBm or more. Therefore, how to improve the linearity of the device when the output current is small, i.e., when the PHEMT is just turned on, is currently an urgent technical problem to be solved.
[0036] The following describes technical terms in the embodiments of this application.
[0037] (1)LNA
[0038] LNA is an important circuit in wireless communication systems, it is an amplifier with a very low noise factor, it functions to amplify the signal received by the antenna. The received signal from the antenna is weak. To amplify such a weak signal, it is of utmost importance to ensure the signal quality. For this purpose, the LNA cannot introduce too much noise, otherwise the signal will be further degraded and demodulation cannot be performed.
[0039] Most LNAs use transistors and field effect transistors. In the embodiment of this application, the LNA uses pseudomorphic high electron mobility transistors.(P It uses HEMT.
[0040] (2) Intermodulation signal
[0041] An intermodulation signal is an intermodulation frequency signal that is generated when an unmodulated signal and a useful signal pass through a nonlinear device and interact with each other. The frequency of the intermodulation signal is very close to that of the useful signal, so the back-end filter can hardly suppress the intermodulation signal. As shown in Figure 1, when a received signal (i.e., a useful signal) and a transmitted signal (i.e., an interference signal) leaking into the receiving link pass through the LNA at the same time, due to nonlinear effects, the frequency of the intermodulation signal generated by the two signals can sometimes be exactly equal to or close to the frequency of the received signal, so that the intermodulation signal passes through the receiver smoothly. The third-order intermodulation signal causes the most serious interference, called intermodulation interference.
[0042] (3) PHEMT
[0043] PHEMT is an improvement over high electron mobility transistor (HEMT). PHEMT has a double heterojunction structure, and the two-dimensional electron gas (2DEG) of PHEMT has double limiting effects on both sides of the potential well. Therefore, compared to HEMT, PHEMT has a higher electron surface density and higher electron mobility.
[0044] FIG. 2A shows a typical double δ-doped PHEMT.
[0045] (4) OIP3
[0046] The linearity of a PHEMT is determined by the output third-order intercept point. (OThe OIP3 value can be expressed by the OIP3 (Oinput-output-frequency) of the PHEMT. A higher OIP3 value indicates a higher linearity. The OIP3 value can be obtained by drawing from the PHEMT radio frequency input and output curves. As shown in FIG. 3, specifically, two curves are drawn. One curve is the curve of the amplified signal power at the input frequency versus the input power, and the other curve is the curve of the third-order intermodulation signal versus the input power. In the logarithmic coordinate system, the curve of the linear amplified signal is represented as a straight line with a slope of 1, and the curve of the third-order intermodulation signal is represented as a straight line with a slope of 3. The output signal power corresponding to the intersection of these two curves is the OIP3 value.
[0047] (5) Output current
[0048] The output current in the embodiment of this application is also referred to as the current output from the drain. The small output current is also simply referred to as the small current. The small output current means that the output current is smaller than a first threshold or smaller than a second threshold, and the first threshold or the second threshold can be the current density that exists when the PHEMT is just turned on. Alternatively, the small output current means that the output current density of the PHEMT is 40-200mA / mm or 60-100mA / mm. The first threshold may or may not be equal to the second threshold. In the embodiment of this application, an example in which the small output current is 60mA / mm is described. The current in the embodiment of this application also refers to the current density.
[0049] In the embodiment of this application, the thickness and materials of all layer structures in the PHEMT are designed to adjust the energy band structure of the channel layer of the PHEMT, in particular the thickness of the epitaxial layer, the doping concentration of the doped layer, and other structures. Therefore, when the PHEMT operates under a small output current, the conduction band energy level of the channel layer is lower than the Fermi energy level. As the gate voltage increases, the increase in the electron concentration near the Fermi energy level has less effect on the electron concentration of the channel layer. This improves the linearity of the PHEMT. In terms of the conduction band structure, the conduction band gradient of the channel layer is smaller in the on-state (3).
[0050] The mechanism for adjusting the energy band structure of the channel layer will now be described.
[0051] There are several main factors that affect the energy band structure of the channel layer:
[0052] a: The distance between the channel layer and the gate. This distance affects the energy level of the channel layer.
[0053] b: doping concentration of upper / lower doped layers. The doped layers change the degree of energy band bending of the structure above them. For example, increasing the doping concentration of the upper doped layer increases the degree of energy band bending of the upper barrier layer, and the gradient of the energy level of the upper barrier layer in the thickness direction increases. In another example, increasing the doping concentration of the lower doped layer increases the degree of energy band bending of the channel layer, and the gradient of the energy level of the channel layer in the thickness direction increases. Thus, decreasing the concentration of the lower doped layer decreases the degree of energy band bending of the channel layer. Small The energy level gradient of the channel layer in the thickness direction is Small During operation under low output current, the conduction band energy level of the channel layer becomes lower than the Fermi energy level.
[0054] c: the ratio of doping concentrations of the upper / lower doped layers. To meet the requirement that the conduction band energy level of the channel layer is lower than the Fermi energy level when the PHEMT operates under a small current, the doping concentration of the lower doped layer is reduced. However, the total doping concentration of the upper / lower doped layers affects the carrier concentration, which affects the performance of the device. And a higher doping concentration indicates a higher turn-on voltage of the device. Therefore, the total doping concentration should not be too low or too high.
[0055] d: Thickness of the channel layer. The thickness of the channel layer affects the distribution of electrons and current. A thicker channel layer can avoid the electrons accumulating under the channel layer when the device is operating, resulting in a more uniform distribution of electrons. This improves the linearity of the PHEMT.
[0056] Therefore, with reference to the above factors, the embodiments of this application provide the following solutions for adjusting the energy band structure of the channel layer: Therefore, when the PHEMT operates under a small current, the conduction band energy level of the channel layer is lower than the Fermi energy level.
[0057] Solution 1: By removing the lower doped layer, the gradient of the energy level of the channel layer in the thickness direction becomes smaller, and the conduction band energy level of the channel layer becomes lower than the Fermi energy level during operation under small current.
[0058] Solution 2: If the total doping concentration is not reduced, the doping concentration of the lower doped layer is reduced, or the ratio of the doping concentration of the upper doped layer to the doping concentration of the lower doped layer is increased.
[0059] Solution 3: The thickness of the channel layer is increased.
[0060] There may also be combinations of the above solutions 1, 2 and 3, combinations of adjusting the thickness of the upper barrier layer and / or the thickness of the upper separation layer with at least one of the above solutions 1, 2 and 3, and the like.
[0061] In the above solution, the thickness and doping concentration of all the layer structures in the PHEMT are adjusted so that the conduction band energy level of the channel layer is lower than the Fermi energy level during operation under small current. Thus, the linearity of the PHEMT can be improved, the linearity of the LNA formed by the PHEMT can be improved, the intermodulation signal output by the LNA can be reduced, and the signal quality can be improved.
[0062] Hereinafter, the embodiments of this application will be described clearly and in detail with reference to the accompanying drawings in the embodiments of this application.
[0063] 4 is a schematic diagram of the cross-sectional structure of a PHEMT. The energy band structure of the channel layer is adjusted by using the above solution 1 or by using a combination of the above solution 1 and at least one of adjusting the thickness of the upper barrier layer and / or the thickness of the upper isolation layer, the above solution 2, and the above solution 3.
[0064] As shown in FIG. 4, the PHEMT is single δ-doped and includes a substrate 1, a buffer layer 2, a lower barrier layer 3, a channel layer 4, an upper isolation layer 5, a first doped layer 6, an upper barrier layer 7, and a cap layer 8, which are arranged in this order on the substrate 1, a source 9 and a drain 10 arranged on the cap layer 8, and a gate electrode 11 arranged on the upper barrier layer 7. T1 1. In the following, the function, composition, thickness, etc. of each layer will be described using a GaAs-based PHEMT as an example.
[0065] The substrate 1 may be a gallium arsenide (GaAs) wafer.
[0066] The buffer layer 2 can be made of gallium arsenide (GaAs) or aluminum gallium arsenide (Al x Ga 1-x As), provided that 0 < x < 1. The buffer layer 2 can prevent defects in the substrate 1 from entering the core of the PHEMT. Generally, the Al doping concentration in the buffer layer can be gradually increased in the direction away from the substrate 1 so as to reduce the defects caused by the lattice mismatch between the substrate 1 and the lower barrier layer 3.
[0067] The lower barrier layer 3 can be made of aluminum gallium arsenide (Al x Ga 1-x As). The bandgap width of the lower barrier layer 3 is larger than that of the channel layer 4. The lower barrier layer 3 and the channel layer 4 form a heterojunction that is a negative junction. The thickness of the lower barrier layer 3 can be 10 - 50 nm or 10 - 25 nm, and can be, for example, 17 nm or 20 nm, etc.
[0068] For example, the maximum Al doping concentration in the buffer layer 2 is less than or equal to the Al doping concentration in the lower barrier layer 3.
[0069] The channel layer 4 can be made of indium gallium arsenide (In y Ga 1-y As), provided that 0 < y < 1. For example, the range of the value of y can be 0.1 - 0.5. For example, y is 0.22 or 0.3.
[0070] In some embodiments, the thickness of the channel layer 4 can be greater than 8 nm or 12 nm and less than the critical thickness of the channel epitaxial layer. For example, In 0.22 Ga 0.78The critical thickness of the As channel layer is about 20 nm. The critical thickness increases as the indium content decreases. Optionally, the thickness of the channel layer 4 can be 8-30 nm, 12-30 nm, or 15-20 nm, such as 17 nm, 18 nm, 20 nm, or 24 nm. Increasing the thickness of the channel layer 4 can improve the distribution of electrons and current in the channel layer 4. This avoids the accumulation of electrons under the channel layer 4, reduces the current in the vertical direction (i.e., the stacking direction of the PHEMT), and improves the linearity of the PHEMT.
[0071] For example, the channel layer 4 is In 0.22 Ga 0.78 It is made of As and has a thickness of 18 nm.
[0072] The upper isolation layer 5 is configured to separate the channel layer 4 from the first doped layer 6 to prevent doping impurities of the first doped layer 6 from entering the channel layer 4. The upper isolation layer 5 may be made of aluminum gallium arsenide and may have a thickness of 2-6 nm, for example 4 nm or 6 nm. The upper isolation layer 5 is also referred to as the first isolation layer.
[0073] The first doped layer 6 can be δ-doped and is also referred to as the upper δ-doped layer. The first doped layer 6 can be silicon doped. A thin layer of silicon is grown on the upper isolation layer 5 as a doping impurity, which provides a two-dimensional electron gas after being ionized. The first doped layer 6 can be a few atoms thick, with a thickness of less than 2 nm, for example, 1 nm or less.
[0074] For example, the doping concentration of the first doped layer 6 is the sum of the doping concentrations of the two doped layers in a prior art double δ-doped PHEMT, or the doping concentration is 3e12 cm -2 From 5e12cm -2 , 5e12cm -2 From 6e12cm -2 , 1e12cm -2 From 3e12cm -2, 4.6e12cm -2 From 5.5e12cm -2 , or 5.5e12cm -2 From 6.5e12cm -2 For example, 4.5e12cm -2 It is.
[0075] The upper barrier layer 7 is made of aluminum gallium arsenide (Al x Ga 1-x The upper barrier layer 7 may be made of ZnO, As. The band gap width of the upper barrier layer 7 is larger than the band gap width of the channel layer 4. The upper barrier layer 7, the first doped layer 6, the upper separation layer 5, and the channel layer 4 form a heterojunction, which is a positive junction. The thickness of the upper barrier layer 7 may be 10-30 nm or 10-25 nm, for example, 15 nm, 17 nm, or 20 nm.
[0076] The cap layer 8 is made of highly doped gallium arsenide (n + -GaAs) and may have a thickness of 5-10 nm. The cap layer 8 is configured to provide an ohmic contact.
[0077] The cap layer 8 has a through hole. The gate 11 is disposed in the through hole and is not in contact with the cap layer 8. The source 9 and the drain 10 are both disposed on the side of the cap layer 8 remote from the upper barrier layer 7, and are located on either side of the through hole, respectively.
[0078] The source 9, drain 10, and gate 11 are all made of conductive metal. The gate 11 is configured to provide a gate voltage to the PHEMT. When the gate voltage is higher than the turn-on voltage, the source 9 and drain 10 are connected and the drain current is output.
[0079] In this embodiment of the present application, the lower barrier layer 3 is directly connected to the channel layer 4. Here, "direct connection" means direct contact. In other words, no other layer structure is included between the lower barrier layer 3 and the channel layer 4.
[0080] The buffer layer 2, the lower barrier layer 3, the upper isolation layer 5, and the upper barrier layer 7 are all made of aluminum gallium arsenide (Al x Ga 1-x As) may be used, but the doping concentration of Al in these layers may be the same or different, and this is not a limitation here. For example, the lower barrier layer 3, the upper isolation layer 5, and the upper barrier layer 7 may all be made of Al. 0.22 Ga 0.78 Use As.
[0081] According to the single δ-doped PHEMT provided in this embodiment of the application, the band gap width of the upper barrier layer 7 and the band gap width of the lower barrier layer 3 are both larger than the band gap width of the channel layer 4, so the single δ-doped PHEMT has one doped positive junction (upper barrier layer 7 / first doped layer 6 / upper isolation layer 5 / channel layer 4) and one undoped negative junction (channel layer 4 / lower barrier layer 3). When the gate voltage is higher than the turn-on voltage, the doping impurities are ionized. Because the band gap width of the upper barrier layer 7 and the band gap width of the channel layer 4 are different, the conduction band is not continuous, and electrons move to one side of the channel layer to form a two-dimensional electron gas. The source 9 and the drain 10 are connected to output the drain current.
[0082] In this embodiment of the application, based on the above-mentioned adjustment mechanism of the energy band structure of the channel layer, the lower doped layer is removed, the appropriate concentration of the upper doped layer (i.e., the first doped layer 6) is selected, and the appropriate thicknesses of the channel layer 4, the upper barrier layer 7, and the upper isolation layer 5 are selected. In the resulting single δ-doped PHEMT, when the output current is a small current, the conduction band energy level of the channel layer 4 is lower than the Fermi energy level, or the conduction band energy level at the interface between the channel layer 4 and the upper isolation layer 5 is lower than the Fermi energy level. Optionally, in the PHEMT, when the output current is a small current, the conduction band energy level of the channel layer 4 is substantially decreased in the thickness direction. In this embodiment of the application, the thickness direction is the direction from the cap layer to the substrate. See the direction shown in FIG. 4. This substantial decrease may include the following two cases:
[0083] 1. The conduction band energy level of the channel layer 4 decreases in the thickness direction. In other words, a larger distance from the upper barrier layer indicates a lower conduction band energy level.
[0084] 2. The conduction band energy level of the channel layer 4 has a substantial / macroscopically decreasing tendency in the thickness direction. In other words, the conduction band energy level is allowed to be relatively high in a small region microscopically separated from the upper barrier layer.
[0085] For example, the thickness of the upper barrier layer 7 is 3-7 nm. The thickness of the upper isolation layer 5 is 4 nm. The doping concentration of the first doped layer 6 is 3.0-4.5e12 cm -2 The thickness of the channel layer is 12 nm.
[0086] For example, the thickness of the upper barrier layer 7 is 5 nm. The thickness of the upper isolation layer 5 is 3-5 nm. The doping concentration of the first doped layer 6 is 3.0-4.5e12 cm -2 The thickness of the channel layer is 8-14 nm.
[0087] Different from the double δ-doped PHEMT in the prior art, this embodiment of the application provides a single δ-doped PHEMT. If the total doping concentration remains unchanged or is not reduced, the δ-doping is performed only on the upper barrier layer 7, and the lower doped layer is removed to reduce the gradient of the conduction band energy level of the channel layer 4 in the thickness direction, which can improve the linearity of the PHEMT under small output current without affecting the gain (i.e., transconductance gm).
[0088] In addition, the thickness of the channel layer 4 is increased, thereby improving the distribution of electrons and current in the channel layer 4. This can prevent accumulation of electrons under the channel layer 4, reduce the current in the vertical direction (i.e., the stacking direction of the PHEMT), and further improve the linearity of the PHEMT.
[0089] 5 shows another PHEMT according to an embodiment of the present application. The PHEMT may be a double-doped PHEMT. In addition to the layer structure shown in FIG. 4, the PHEMT may further include a second doped layer 12 and a lower isolation layer 13.
[0090] Both the first doped layer 6 and the second doped layer 12 may be δ-doped. Thus, the first doped layer 6 is also referred to as the upper δ-doped layer 6, and the second doped layer 12 is also referred to as the lower δ-doped layer 12. Both the first doped layer 6 and the second doped layer 12 may be silicon-doped. A thin layer of silicon is grown on each of the upper isolation layer 5 and the lower barrier layer 3 as a doping impurity, which provides a two-dimensional electron gas after being ionized. Both the first doped layer 6 and the second doped layer 12 may be a few atoms thick, with a thickness of less than 2 nm, for example, 1 nm or less. The doping concentration of the first doped layer 6 is higher than the doping concentration of the second doped layer 12.
[0091] The lower isolation layer 13 is configured to separate the second doped layer 12 from the channel layer 4 to prevent doping impurities of the second doped layer 12 from entering the channel layer 4. The lower isolation layer 13 may be made of gallium arsenide and may have a thickness of 2-6 nm, for example 4 nm. The lower isolation layer 13 is also referred to as the second isolation layer.
[0092] It should be noted that the buffer layer 2, the lower barrier layer 3, the upper isolation layer 5, the lower isolation layer 13, and the upper barrier layer 7 may all be made of aluminum gallium arsenide, but the doping concentrations of Al in these layers may be the same or different, and this is not a limitation here.
[0093] According to the double δ-doped PHEMT provided in this embodiment of the application, the band gap width of the upper barrier layer 7 and the band gap width of the lower barrier layer 3 are both larger than the band gap width of the channel layer 4, so the double δ-doped PHEMT has one doped positive junction (upper barrier layer 7 / first doped layer 6 / upper isolation layer 5 / channel layer 4) and one doped negative junction (channel layer 4 / lower isolation layer 13 / second doped layer 12 / lower barrier layer 3). When the gate voltage is higher than the turn-on voltage, the doping impurities are ionized. Because the band gap width of the barrier layer and the band gap width of the channel layer 4 are different, the conduction band is not continuous, and electrons move to one side of the channel layer to form a two-dimensional electron gas. The source 9 and the drain 10 are connected to output the drain current.
[0094] In some embodiments, the energy band structure of the channel layer is adjusted by using solution 3 described above. In the PHEMT shown in FIG. 5, the thickness of the channel layer 4 is greater than 12 nm and less than the critical thickness of the channel epitaxial layer (i.e., does not exceed the movement limit of the InGaAs channel epitaxial layer). Therefore, in the double δ-doped PHEMT, when the output current is a small current, the conduction band energy level of the channel layer 4 is lower than the Fermi energy level. For example, the thickness of the channel layer 4 is 8-30 nm, 12-30 nm, or 15-20 nm, such as 17 nm, 18 nm, 20 nm, or 24 nm. The thickness or critical thickness of the channel layer 4 is related to the indium content. A lower indium content indicates a higher thickness. For example, the channel layer 4 is made of InGaAs. 0.22 Ga 0.78 It is made of As and has a thickness of 18 nm.
[0095] In the above-mentioned double δ-doped PHEMT, the thickness of the channel layer 4 is increased, which can improve the distribution of electrons and current in the channel layer 4. This can avoid the accumulation of electrons under the channel layer 4, reduce the current in the thickness direction, and improve the linearity of the PHEMT.
[0096] In some other embodiments, the energy band structure of the channel layer is adjusted by using the above-mentioned solution 2. Thus, in the double δ-doped PHEMT, when the output current is small, the conduction band energy level of the channel layer 4 is lower than the Fermi energy level.
[0097] For example, the doping concentration of the first doped layer 6 is 3.5-4.5e12 cm -2 The doping concentration of the second doped layer 12 is 3-5e11 cm -2 It is.
[0098] For example, the ratio of the doping concentration of the first doped layer 6 to the doping concentration of the second doped layer 12 is greater than a preset value. The preset value is 6 or more, such as 9, 10, 15, 30, 70, 100, or 150.
[0099] In another example, the doping concentration of the first doped layer 6 is 4-6e12 cm -2 The doping concentration of the second doped layer 12 is 2e8cm -2 From 3e11cm -2 , 1e6cm -2 From 1e11cm -2 , or 1e6cm -2 From 1e8cm -2 It is.
[0100] In another example, the doping concentration of the first doped layer 6 is 3.5-4.5e12 cm -2 , 4.5-6e12cm -2 , or 4.6-5.5e12cm -2 The doping concentration of the second doped layer 12 is 2e8cm -2 From 3e11cm -2 It is.
[0101] In the double δ-doped PHEMT, the doping concentration and the ratio of the doping concentrations of the upper / lower δ-doped layers are adjusted to reduce the gradient of the energy level of the channel layer 4 caused by the second doped layer 12. Thus, when the device operates under low current, the conduction band energy level of the channel layer 4 is lower than the Fermi energy level, which improves the linearity of the device.
[0102] In some other embodiments, the energy band structure of the channel layer is adjusted by using the above-mentioned solution 2 and solution 3. Therefore, in the double δ-doped PHEMT, when the output current is small, the conduction band energy level of the channel layer 4 is lower than the Fermi energy level. For example, the thickness of the channel layer is 18 nm. The doping concentration of the first doped layer 6 is 3.5-4.5e12 cm-2 The doping concentration of the second doped layer 12 is 2e8cm -2 From 3e11cm -2 Alternatively, the ratio of the doping concentration of the first doped layer 6 to the doping concentration of the second doped layer 12 is set to be greater than a preset value. The preset value is 6 or more, for example, 9.
[0103] In the double δ-doped PHEMT, the doping concentration and the ratio of the doping concentrations of the upper / lower doped layers are adjusted to reduce the energy level gradient of the channel layer 4 caused by the lower doped layer, so that the conduction band energy level of the channel layer 4 is lower than the Fermi energy level when the device operates under low current, which improves the linearity of the device.
[0104] In some other embodiments, the energy band structure of the channel layer is adjusted by using a combination of solution 2 and / or solution 3 and adjusting the thickness of the upper barrier layer and / or the upper isolation layer, so that in the double δ-doped PHEMT, when the output current is small, the conduction band energy level of the channel layer 4 is lower than the Fermi energy level.
[0105] For example, the thickness of the upper barrier layer 7 is 3-7 nm. The thickness of the upper isolation layer 5 is 3-5 nm. The doping concentration of the first doped layer 6 is 3.0-4.5e12 cm -2 The doping concentration of the second doped layer 12 is 3-5e11 cm -2 The thickness of the channel layer is 18 nm.
[0106] In another example, the thickness of the upper barrier layer 7 is 3-7 nm. The thickness of the upper isolation layer 5 is 3-5 nm. The doping concentration of the first doped layer 6 is 3.0-4.5e12 cm -2 The doping concentration of the second doped layer is 2e8cm -2 From 3e11cm -2 The thickness of the channel layer is 14-20 nm.
[0107] In another example, the thickness of the upper barrier layer 7 is 5 nm. The thickness of the upper isolation layer 5 is 4 nm. The doping concentration of the first doped layer 6 is 3.0-4.5e12 cm -2 The doping concentration of the second doped layer is 2e8cm -2 From 3e11cm -2 The thickness of the channel layer is 14-20 nm.
[0108] Also, in the above embodiment, when the PHEMT is under a small output current, the conduction band energy level of the channel layer 4 is substantially lowered in the thickness direction.
[0109] 4 or 5 may instead be made of other materials and have other thicknesses, such as ternary, quaternary, or multi-nary compounds of gallium arsenide, hi another example, the PHEMT is a gallium phosphide (GaP) based PHEMT.
[0110] It should be noted that in the description of FIGS. 4 and 5, the upper doped layer or the upper δ-doped layer corresponds to the first doped layer 6 , and the lower doped layer or the lower δ-doped layer corresponds to the second doped layer 12 .
[0111] It should also be noted that for clarity of illustration, the layer thicknesses in FIGS. 4 and 5 do not correspond to the actual thicknesses or thickness ratios of the layers, and specific thickness ranges are subject to the thicknesses specified for the layers described above.
[0112] Compared with the prior art, the OIP3 value of the single δ-doped PHEMT provided in this embodiment of this application is significantly higher. When the output drain current is 60-100mA, the OIP3 value of the single δ-doped PHEMT increases by more than 5dBm. Specific data is shown in FIG. A and Figure 6B As shown in Figure 6 A and Figure 6B1 compares the simulation result of the OIP3 value of the double δ-doped PHEMT in the prior art with the simulation result of the OIP3 value of the single δ-doped PHEMT provided in this embodiment of the application. It can be seen that, compared with the double δ-doped PHEMT in the prior art, the OIP3 value of the single δ-doped PHEMT provided in this embodiment of the application is increased by more than 5 dBm when the output current is 50-200 mA.
[0113] With reference to FIGS. 7A-7C, the principle of improving the linearity of PHEMTs through single δ-doping is explained below.
[0114] FIG. 7A is an energy band diagram of a double δ-doped PHEMT in the prior art. FIG. 7B is an energy band diagram of a single δ-doped PHEMT according to an embodiment of this application. FIG. 7C compares the conduction band structure of a double δ-doped PHEMT with that of a single δ-doped PHEMT in the channel layer. It should be understood that the Fermi energy level increases as the gate voltage increases. FIG. 7A-FIG. 7C all show the position of the Fermi energy level when the drain current (i.e., the output current) is 60 mA.
[0115] Due to the concentrated distribution of space charges at the δ-doping position, the energy band on the left side of the δ-doping position is bent. It can be seen from the energy band structures that when the output current is 60 mA, the conduction band structure of the channel layer of the single δ-doped PHEMT is close to a rectangle, while that of the double δ-doped PHEMT is close to a triangle, as shown by the shaded area in Fig. 7C. Since the electron concentration in the channel layer is proportional to the area of the polygon enclosed by the conduction band and the Fermi energy level, the rectangular conduction band structure has a higher linearity than the triangular conduction band structure in the process of changing the gate voltage.
[0116] All the electrons in the channel layer of the single δ-doped PHEMT are provided by the top δ-doped layer. Therefore, to obtain the same output current, the conduction band of the channel layer of the single δ-doped PHEMT is further below the Fermi energy level compared to the conduction band of the channel layer of the double δ-doped PHEMT. A Fermi-Dirac distribution occurs for electrons near the Fermi energy level. Specifically, the probability of electron generation near the Fermi energy level (shown by the area surrounded by the dashed line in FIG. 7C) is 0.5, and the probability of electron generation below the Fermi energy level is 1. Thus, the probability of electron generation near the Fermi energy level is nonlinear to some extent when the gate voltage is changed. However, the area occupied by electrons near the Fermi energy level has a small proportion in the overall conduction band structure of the channel layer of the single δ-doped PHEMT. Therefore, the single δ-doping design, whose conduction band is further below the Fermi energy level, can provide higher linearity.
[0117] With reference to FIG. 8, the principle of improving the linearity of a PHEMT by increasing the thickness of the channel layer will be explained below.
[0118] FIG. 8 is a diagram of the electron concentration distribution of the double δ-doped PHEMTs with channel layers of 12 nm and 18 nm, respectively, under an output current of 60 mA / mm. It can be seen from FIG. 8 that, when the PHEMT operates under an output current of 60 mA / mm, in the PHEMT with a channel layer of 12 nm, a large amount of electrons are collected under the channel layer, and the current in the channel layer has a large z-direction component. In comparison, in the PHEMT with a channel layer of 18 nm, the electrons are more uniformly distributed in the channel layer when the PHEMT operates. This reduces the z-direction component of the current in the channel layer, reduces the complexity of controlling the channel current with a gate, and improves the linearity of the PHEMT.
[0119] The application scenarios of the PHEMT provided in the embodiments of this application are described below.
[0120] An embodiment of this application further provides an LNA. As shown in FIG. 9, the LNA includes an input matching network, a bias circuit, a PHEMT, and an output matching network. The PHEMT can be the PHEMT shown in FIG. 4 or FIG. 5. The bias circuit is configured to provide a bias voltage for proper operation of the PHEMT, i.e., to set the gate, source, and drain of the PHEMT to a required potential. The input matching network is configured to realize a match between the output impedance of the signal source and the input impedance of the LNA, so that the LNA obtains the maximum excitation power. The output matching network is configured to transform an external load resistance into an optimal load resistance used by the amplifier to ensure the maximum output power.
[0121] An embodiment of the present application further provides a receiver or transceiver. The receiver or transceiver may include an LNA as shown in FIG. 9. Optionally, the receiver or transceiver may further include a duplexer, a bandpass filter, a digital-to-analog converter ( DAC ), and the like.
[0122] The LNA can be applied to a radio frequency system. As shown in Fig. 10, the radio frequency system can be divided into a transmission link and a reception link. In the application scenario of this application, the low noise amplifier LNA in the reception link of the radio frequency system can be the LNA shown in Fig. 9, and is configured to amplify the signal received by the antenna.
[0123] The transmission link consists of a power amplifier (PA), Bah, At least one filter Ta, At least one mixi sa, At least one local oscillator (L O), and at least one amplifier. (AThe receive link may include a low noise amplifier (LNA), at least one filter, at least one mixer, at least one local oscillator (LOO), etc. (L O), and at least one amplifier. The at least one filter may include an image rejection filter. Ta, It may include an intermediate frequency filter (IF filter), or other filters.
[0124] It should be understood that Fig. 10 is merely an illustrative example, and the radio frequency system may alternatively be of other circuit configurations and may alternatively include fewer components than those of Fig. 10, and this is not intended to be limiting here.
[0125] The application further provides a radio frequency circuit, including a PHEMT as shown in Fig. 4 or Fig. 5, or including an LNA as shown in Fig. 9, adapted for wireless communication field, and configured to process a signal received by an antenna and / or control the antenna to transmit a signal.
[0126] The application further provides a radio frequency chip, which is configured to process a signal received by a receive antenna, send the signal to a processor, receive instructions from the processor, and control a transmit antenna to transmit a signal.
[0127] The application further provides an electronic device, which may be a device with a radio frequency function or a wireless communication function, such as a mobile phone, a tablet computer, an e-book reader, a television, a notebook computer, a digital camera, an in-vehicle device, a wearable device, a base station, or a router, etc. The electronic device may include at least one of the above-mentioned PHEMT, the LNA, the radio frequency system, the radio frequency circuit, and the radio frequency chip.
[0128] In short, the above embodiments are merely intended to describe the technical solutions of this application, and are not intended to limit this application. Although this application has been described in detail with reference to the above embodiments, those skilled in the art should understand that without departing from the scope of the technical solutions of the embodiments of this application, modifications to the technical solutions recorded in the above embodiments, or equivalent substitutions to some technical features thereof, can still be made.
Claims
1. A pseudomorphic high electron mobility transistor (PHEMT), comprising: A channel layer; a lower barrier layer and an upper barrier layer disposed on either side of the channel layer, the lower barrier layer being connected to the channel layer; a first isolation layer and a first doped layer disposed between the channel layer and the upper barrier layer, the first isolation layer configured to isolate the first doped layer from the channel layer, the first doped layer configured to provide a two-dimensional electron gas; having When an output current of the PHEMT is smaller than a first threshold, a conduction band energy level of the channel layer is lower than a Fermi energy level. PHEMT.
2. The PHEMT of claim 1 , wherein the lower barrier layer is directly connected to the channel layer.
3. The first doped layer is silicon doped and has a doping concentration of 3e12 cm -2 From 5e12cm -2 The PHEMT of claim 2 ,
4. 2. The PHEMT of claim 1, wherein the lower barrier layer is connected to the channel layer through a second isolation layer and a second doped layer, the second isolation layer configured to isolate the channel layer from the second doped layer, and the second doped layer configured to provide a two-dimensional electron gas.
5. The doping concentration of the first doped layer is 3.5e12 cm -2 From 4.5e12cm -2 and the doping concentration of the second doped layer is 3e11 cm -2 From 5e11cm -2 The PHEMT of claim 4 ,
6. The PHEMT of claim 4 , wherein a ratio of a doping concentration of the first doped layer to a doping concentration of the second doped layer is greater than a preset value.
7. The PHEMT of claim 6 , wherein the preset value is greater than nine.
8. 8. The PHEMT of claim 3 , wherein the values of the concentration of the first doped layer and the concentration of the second doped layer enable the conduction band energy level of the channel layer to be lower than the Fermi energy level when the PHEMT is in an on-state.
9. The PHEMT according to any one of claims 1 to 8, wherein the channel layer has a thickness of 15-20 nm.
10. 10. The PHEMT of claim 1, wherein the conduction band energy level of the channel layer is substantially decreased in a thickness direction when the output current of the PHEMT is less than a second threshold.
11. 11. The PHEMT of claim 1, further comprising a cap layer, a source, a drain, and a gate, the cap layer being disposed on a side of the upper barrier layer away from the channel layer and having a through hole for providing an ohmic contact, the gate being disposed within the through hole, and the source and drain being both disposed on a side of the cap layer away from the upper barrier layer and located on either side of the through hole.
12. 12. The PHEMT of claim 1, wherein the channel layer is made of indium gallium arsenide and the upper barrier layer, the lower barrier layer, or the separation layer is made of aluminum gallium arsenide.
13. A low noise amplifier comprising a PHEMT according to any one of claims 1 to 12.
14. A radio frequency circuit comprising the low noise amplifier of claim 13.
15. A radio frequency chip comprising at least one of the PHEMT according to any one of claims 1 to 12, the low noise amplifier according to claim 13, and the radio frequency circuit according to claim 14.
16. 16. An electronic device comprising at least one of the PHEMT according to claim 1, the low noise amplifier according to claim 13, the radio frequency circuit according to claim 14, and the radio frequency chip according to claim 15.
Citation Information
Patent Citations
Linear heterojunction field-effect transistor
JP1994168963A
Semiconductor device
JP1999068086A
Single-voltage supplied pseudomorphic high-electron-mobility transistor (PHEMT) power device and method for manufacturing the same
JP2010506397A
Single voltage supply pseudomorphic high electron mobility transistor (PHEMT) power device and process for manufacturing the same
US20100102358A1
Semiconductor device, antenna switch circuit, and wireless communication apparatus
US20170330880A1
