A method for characterizing trap parameters of GaN HEMT devices using microsecond transient curves

Abstract

The present invention discloses a method for characterizing the trap parameters of a GaN HEMT device using a microsecond transient curve, and relates to the field of semiconductor device reliability. The microsecond transient voltage curve of the GaN HEMT device under a constant electrical bias is collected through circuit design, and the relevant trap information is obtained by further analysis and processing using a structure function method. The method mainly comprises: placing the device under test on a constant temperature platform and connecting it to a trap test circuit; applying a bias voltage to the device through the circuit during the trap filling process; shortening the switching time to the microsecond level through a fast switch during the trap release process, and simultaneously collecting the changes in the transient voltage curve and analyzing and processing it through software to obtain information related to the internal traps of the device. The present invention improves the collection of transient voltage curves to the microsecond level through circuit design, and can be used for trap testing of GaN HEMT devices produced by different manufacturers, with good versatility.

Description

Technical field:

[0001] The present invention relates to the field of semiconductor device reliability and is mainly used for measuring and characterizing internal trap parameters of GaNHEMT (Gallium Nitride-Based High-electron-mobility Transistor) devices. Background technology:

[0002] The development of wide-bandgap semiconductor devices, represented by GaN HEMT devices, has become a significant driving force for the continued advancement of power electronics. GaN HEMT devices, characterized by high electron mobility and a high breakdown electric field, hold broad promise for high-frequency, high-power applications. Compared to Si- and GaAs-based devices, GaN-based devices exhibit degradation and failure phenomena whose mechanisms are not yet fully understood. These phenomena are closely related to internal device traps, and the resulting instability in electrical parameters poses a potential risk to system reliability.

[0003] Currently, the trap effect in GaN HEMT devices is primarily manifested through phenomena such as gate delay, drain delay, and current collapse. The degradation of the device's drain-source current or drain-source voltage over time directly reflects the impact of traps on the device's electrical performance. Under large gate-source and drain-source voltage biases, the device's internal traps are filled with electrons. Subsequently, under a constant drain-source current bias, electrons are released from the traps, increasing the two-dimensional electron gas concentration and decreasing the transient drain-source voltage. However, testing using semiconductor parameter analyzers typically uses millisecond-scale time as the starting point to capture transient curve changes, making it impossible to achieve faster switching speeds to test and capture microsecond-scale transient responses. GaN HEMT devices are widely used in high-frequency applications. The transient curve changes caused by the filling of internal traps in the device follow an exponential relationship. These microsecond-scale transient curve changes contain a large amount of trap information, making obtaining microsecond-scale trap information crucial for further exploring the trapping mechanism.

[0004] This technology proposes a method for characterizing trap parameters in GaN HEMT devices by collecting microsecond-scale transient curves. This technique can be applied to extract parameters of traps within GaN HEMT devices, achieving faster switching speeds and enabling the collection of microsecond-scale transient curves during the trap release process. This allows for more complete trap information and further study of its mechanism of action. This method is highly versatile and can be applied to trap testing in various GaN HEMT device models. Summary of the invention:

[0005] This invention discloses a method for characterizing the trap parameters of GaN HEMT devices using microsecond-scale transient curves. Through circuit design, the microsecond-scale transient voltage curve of the GaN HEMT device is collected under a constant electrical bias. The structure function method is then used to further analyze and process the relevant trap information. The main inventive features of this method are: a fast switching circuit for trap testing is designed to enable testing of microsecond-scale transient curves and the collection of corresponding trap information; and the stable application of a 48V drain-source voltage and a -10V gate-source voltage bias during the trap-filling phase is achieved, meeting the voltage bias requirements of GaN HEMT devices during the trap-filling phase.

[0006] A method for characterizing trap parameters of a GaN HEMT device using microsecond transient curves, characterized by:

[0007] 1. The device under test 109 is fixed on a constant temperature platform 110 with a temperature of T1, and is connected to the constant current source circuit 106, the leakage voltage control circuit 107, the gate voltage control circuit 108 and the amplifier circuit 105; wherein the leakage voltage control circuit 107 is connected to the device under test 109 through a diode load 111, and the gate voltage control circuit 108 is connected to the device under test through a bipolar transistor 112. The computer 101 realizes real-time acquisition of transient voltage signals through the AD acquisition circuit 103, and realizes voltage signal amplification and noise reduction processing through the amplifier circuit 105 and the signal-to-noise ratio enhancement circuit 104. The amplifier circuit 105 can amplify the collected voltage signal by 10 times, and the sampling speed of the AD acquisition circuit is 1MHz. The computer 101 sets the gate-source voltage V in the filling stage through the FPGA control module 102 GF , drain-source voltage V DF , filling time t1, and the drain-source current I during the test phase DM , test time t2, the trap filling and testing process starts under the control of computer 101.

[0008] 2. In the trap filling stage, the computer 101 controls the drain voltage control circuit 107 and the gate voltage control circuit 108 through the FPGA control module 102 to apply a negative gate-source voltage V GF and positive drain-source voltage V DF , used to fill the internal traps of the device, where the negative gate-source voltage V GF The range is -10V to 0V, the positive drain-source voltage V DF The range is 0V to 48V, and the filling time t1 ranges from 1ms to 30s.

[0009] 3. After the filling time t1 is completed, the computer 101 controls the bipolar transistor 112 through the FPGA control module 102 to realize the rapid switching of the gate voltage control circuit 108 for the negative gate-source voltage, wherein the switching range of the gate-source voltage can reach -10V to 0V; at the same time, the diode load 111 is used to reduce the establishment time of the constant current source during the switching process, thereby realizing the switching of the leakage voltage control circuit 107 and the constant current source circuit 106, wherein the switching time can be controlled within 10μs, thereby obtaining the microsecond transient curve change. After the circuit switching, the trap release process is tested. A constant drain-source current is applied to the device through the constant current source circuit 106 at a gate-source voltage of 0V, and the drain-source voltage of the device under test is obtained through the AD acquisition circuit 103. The test time is t2. During the test, the 0V gate-source voltage can ensure that the device under test is in the on state (the threshold voltage of the depletion-type device is less than 0V), and the potential between the gate and source of the device is the same, thereby avoiding the influence of the external gate-source electric field; the drain-source current I applied to the device DM The range is 0~300mA, the transient curve sampling accuracy can reach 1μs, and the test time t2 range is 10μs~300s.

[0010] 4. The transient voltage data obtained in step 3 is amplified and de-noised by amplifier circuit 105 and signal-to-noise ratio enhancement circuit 104, respectively, to obtain a transient voltage curve with microsecond-to-second variations. The data is then returned to computer 101 for structure function processing to obtain a corresponding time constant spectrum, the abscissa of which represents the time constant corresponding to the trap. The horizontal axis of the time constant spectrum (TCS) corresponds to the test time t2 of the transient voltage response curve, expressed in logarithmic form; the vertical axis represents the relative amplitude obtained after structure function processing. The number of peaks obtained from the time constant spectrum (TCS) represents the number of traps, and the abscissa corresponding to the peak value represents the time constant of the trap.

[0011] The present invention proposes for the first time the use of microsecond transient curves to characterize the trap parameters of GaN HEMT devices. This method can obtain transient response curves from microseconds to seconds, from which more complete trap information can be extracted, thereby effectively characterizing the trap distribution inside the device. Description of the drawings:

[0012] Figure 1 This is a schematic diagram of the trap test circuit;

[0013] Among them, 101: computer; 102: FPGA control module; 103: AD acquisition circuit; 104: signal-to-noise ratio enhancement circuit; 105: amplifier circuit; 106: constant current source circuit; 107: leakage voltage control circuit; 108: gate voltage control circuit; 109: device under test; 110: constant temperature platform; 111: diode load; 112: bipolar transistor;

[0014] Figure 2 It is a microsecond level transient voltage curve test timing diagram;

[0015] Figure 3 It is the transient voltage change curve and time constant spectrum. Specific implementation method:

[0016] The following is an example of using microsecond-level transient curves to characterize the trap parameters of a GaN HEMT device. The commercial GaN HEMT device CGH40010 produced by CREE is selected as the device under test 109.

[0017] 1. The device under test 109 is fixed on a constant temperature platform 110 at a temperature of 298K and connected to the constant current source circuit 106, the leakage voltage control circuit 107, the gate voltage control circuit 108 and the amplifier circuit 105; wherein the leakage voltage control circuit 107 is connected to the device under test 109 through the diode load 111, and the gate voltage control circuit 108 is connected to the device under test through the bipolar transistor 112. The computer 101 realizes real-time acquisition of transient voltage signals through the AD acquisition circuit 103, and realizes voltage signal amplification and noise reduction processing through the amplifier circuit 105 and the signal-to-noise ratio enhancement circuit 104. The computer 101 sets the gate-source voltage V in the filling phase through the FPGA control module 102 GF , drain-source voltage V DF , filling time t1, and the drain-source current I during the test phase DM At test time t2, the computer 101 controls the start of the trap filling and testing process. The test circuit diagram is shown in FIG. Figure 1 shown.

[0018] 2. In the trap filling stage, the computer 101 controls the drain voltage control circuit 107 and the gate voltage control circuit 108 through the FPGA control module 102 to apply a -10V gate-source voltage V GF and 10V drain-source voltage V DF , the filling time t1 is 30s.

[0019] 3. After the filling process is completed, the computer 101 controls the bipolar transistor 112 through the FPGA control module 102 to achieve rapid switching of the gate voltage control circuit 108 between -10V and 0V. At the same time, the diode load 111 is used to reduce the constant current source settling time during the switching process, achieving the switching between the leakage voltage control circuit 107 and the constant current source circuit 106. The switching time can be controlled within 10μs. After the circuit switching, the trap release process is tested. At 0V gate-source voltage, a constant drain-source current I of 200mA is applied to the device through the constant current source circuit 106. DM The curve of the drain-source voltage of the device under test changing with time is obtained by AD acquisition circuit 103. The test time t2 is 120s. Figure 2 shown.

[0020] 4. The transient voltage data obtained in step 3 is amplified and noise-reduced by the amplifier circuit 105 and the signal-to-noise ratio enhancement circuit 104, respectively. The final transient voltage curve is as follows: Figure 3 As shown in (a), a transient voltage change curve within a time range of 10 μs to 120 s is obtained. The transient curve data is returned to the computer 101 for structure function processing to obtain the corresponding time constant spectrum as shown in FIG. Figure 3 (b) The horizontal axis of the time constant spectrum (TCS) corresponds to the test time of the transient curve (120s) and is expressed in logarithmic coordinates; the vertical axis corresponds to the relative amplitude after structure function processing. Figure 3 There are two peaks in (b), indicating that there are two traps in the voltage curve that affect its transient changes; the horizontal axes corresponding to the two peaks are the time constants of the two traps, which are named DP1 and DP2 in order from small to large.

Claims

1. A method for characterizing the trap parameters of a GaN HEMT device using microsecond transient curves, characterized by: 1) The device under test is placed on a constant temperature platform at a temperature of T1 and connected to a constant current source circuit, a drain voltage control circuit, a gate voltage control circuit, and an amplifier circuit; the drain voltage control circuit is connected to the device under test through a diode load, and the gate voltage control circuit is connected to the device under test through a bipolar transistor; the computer uses an AD acquisition circuit to acquire transient voltage signals in real time, and uses an amplifier circuit and a signal-to-noise ratio enhancement circuit to amplify and reduce noise, respectively; the gate-source voltage V in the filling phase is set by the computer. GF , drain-source voltage V DF , filling time t1, and the drain-source current I during the test phase DM ,Test time t2, the trap filling and testing process begins under computer control; 2) During the filling phase, the computer controls the drain voltage control circuit and the gate voltage control circuit to apply a negative gate-source voltage V with a constant pulse width t1 to the device. GF and positive drain-source voltage V DF To fill the internal traps of the device; 3) After the filling time t1, a fast switching circuit constructed by bipolar transistors is controlled by a computer to achieve fast switching of the gate voltage control circuit. At the same time, a diode is used as a load to reduce the establishment time of the constant current source during the switching process. After the circuit is switched, the trap release process is tested. During the test phase, a 0V gate-source voltage is applied to the device to ensure that the device under test is in the on state. The threshold voltage of the depletion-type device is less than 0 V, and the potential between the gate and source of the device is the same to avoid the influence of the external gate-source electric field. At the same time, a constant drain-source current I is applied. DM , the curve of the drain-source voltage of the device under test changing with time is collected through the AD acquisition circuit, and the test time is t2; 4) The transient voltage data obtained in step 3) is amplified and noise-reduced by an amplifier circuit and a signal-to-noise ratio enhancement circuit, respectively, to obtain a transient voltage response curve; the data is returned to a computer for structure function processing to obtain a corresponding time constant spectrum TCS; The horizontal axis of the time constant spectrum TCS corresponds to the test time t2 of the transient voltage response curve and is expressed in logarithmic coordinates; the vertical axis is the relative amplitude of the trap obtained after processing by the structure function method; the number of peaks obtained from the time constant spectrum TCS is the number of traps, and the horizontal axis corresponding to the peak is the time constant of the trap.

2. The method for characterizing the trap parameters of a GaN HEMT device according to claim 1, wherein: During the trap filling phase, the positive drain-source voltage V DF The range is 0 V to 48 V, and the negative gate-source voltage V GF The range is -10 V to 0 V, and the fill time t1 range is 1ms to 30 s.

3. The method for characterizing the trap parameters of a GaN HEMT device according to claim 1, wherein: The switching time between the trap filling phase and the test phase is controlled within 10 μs, and the gate voltage control circuit realizes the switching of the voltage from -10 V to 0 V.

4. The method for characterizing the trap parameters of a GaN HEMT device according to claim 1, wherein: The switching between the trap filling phase and the test phase is achieved by a fast switching circuit constructed by bipolar transistors and diodes, which provides a constant drain-source current while changing the gate-source voltage and starts to collect changes in transient drain-source voltage.

5. The method for characterizing the trap parameters of a GaN HEMT device according to claim 1, wherein: During the test phase, the drain-source current I DM The range is 0 ~ 300 mA, the transient curve sampling accuracy can reach 1μs, and the test time t2 range is 10μs ~ 300 s.

6. The method for characterizing trap parameters of a GaN HEMT device according to claim 1, wherein: The trap parameter to be characterized is the time constant of the trap, and the characterization method is to collect the transient drain-source voltage curve of the device during the trap test phase.

Images