Semiconductor devices and methods for evaluating devices under test
By introducing a reliability stress circuit into the semiconductor device, the problem of inaccurate simulation of dynamic switching conditions in aging tests is solved, enabling more accurate reliability assessment and flexible aging tests.
Patent Information
- Application Number
- CN202611081732.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2025-07-22
- Filing Date
- 2026-07-21
- Publication Date
- 2026-11-17
AI Technical Summary
Existing semiconductor devices cannot accurately simulate dynamic switching conditions during aging tests, leading to inaccurate reliability assessments and affecting functional correctness and system stability.
Employing a reliable stress circuit, including a frequency generator, frequency divider, signal conditioning circuit, and load circuit, it simulates the stress profile of a real operating environment by generating and applying programmable gate and drain signals, supporting repeatable and controlled aging assessments.
It enables accurate simulation of aging and stress testing on semiconductor wafers or within devices, improving the accuracy and flexibility of reliability assessment and adapting to changes in different aging mechanisms and operating environments.
Smart Images

Figure CN123124752A_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to a semiconductor device and a method for evaluating a device under test formed in a functional region of an integrated circuit. Background Technology
[0002] In semiconductor technology, integrated circuit (IC) devices are increasingly expected to meet high performance and reliability standards across a wide range of applications, including computing, communications, automotive, and industrial systems. As devices continue to scale across advanced technology nodes, long-term reliability has become a key consideration during the design, verification, and qualification phases. One form of reliability assessment is aging testing, which evaluates how a semiconductor device degrades under electrical and / or thermal stress. Device aging can lead to changes in electrical parameters such as threshold voltage, drive current, leakage current, on-resistance, and switching speed. These changes can adversely affect functional correctness, timing margins, and system stability. Summary of the Invention
[0003] This disclosure includes a semiconductor device comprising: a semiconductor wafer including an integrated circuit and scribe lines formed in a functional region of the wafer, each integrated circuit being separated by a separate scribe line; a device under test (DUT) formed in at least one of the scribe lines; and a reliability stress circuit formed in at least one of the scribe lines for applying a stress signal to the DUT for evaluation, the reliability stress circuit comprising: a frequency generator for generating a first signal; a frequency divider coupled to the frequency generator for generating a divided signal; a signal conditioning circuit for receiving a selected one of the divided signals and generating a gate control signal for application to a first terminal of the DUT; and a load circuit coupled to a second terminal of the DUT for applying voltage or current stress according to the gate control signal.
[0004] This disclosure includes a method for evaluating a device under test (DUT) formed in an active region of an integrated circuit, the method comprising the steps of: generating a first signal using a frequency generator; generating a lower frequency signal using a frequency divider coupled to the frequency generator; selecting one of the lower frequency signals and generating a gate control signal using a signal conditioning circuit; applying the gate control signal to the gate terminal of the DUT; and applying voltage or current stress to the current carrying terminal of the DUT using a load circuit, wherein the voltage or current stress is applied according to the gate control signal.
[0005] This disclosure includes a method for evaluating a device under test (DUT) formed in a functional region of an integrated circuit, comprising the steps of: forming an integrated circuit in a plurality of functional regions of a semiconductor wafer, each integrated circuit being defined by a separate scribe line; fabricating a reliability stress circuit and a DUT coupled to the reliability stress circuit in at least one of the scribe lines; generating a first signal from the reliability stress circuit; dividing the first signal into lower frequency signals; selecting one of the lower frequency signals based on a test configuration; generating a gate control signal from the selected lower frequency signal; applying the gate control signal to the gate terminal of the DUT; and applying voltage or current stress to the current carrying terminal of the DUT according to the gate control signal. Attached Figure Description
[0006] The nature of this disclosure is best understood when read in conjunction with the accompanying drawings from the following detailed description:
[0007] Figure 1A This is a schematic block diagram illustrating exemplary devices according to various embodiments of the present disclosure;
[0008] Figure 1B This is a schematic block diagram illustrating exemplary reliability stress circuits according to various embodiments of the present disclosure;
[0009] Figure 2A This is a schematic block diagram / circuit diagram illustrating another exemplary device according to various embodiments of the present disclosure;
[0010] Figure 2B This is a schematic timing diagram illustrating the exemplary relationship between frequency signals according to various embodiments of the present disclosure;
[0011] Figure 2C A schematic timing diagram illustrating an exemplary lower-frequency output signal generated from a high-frequency input signal according to various embodiments of the present disclosure;
[0012] Figure 2D This is a schematic timing diagram illustrating exemplary voltage level shifts according to various embodiments of the present disclosure;
[0013] Figure 2E This is a schematic timing diagram illustrating exemplary gate signals according to various embodiments of the present disclosure;
[0014] Figure 2F This is a timing diagram illustrating the illustrative switching frequency variation in response to different drain voltages according to various embodiments of the present disclosure;
[0015] Figure 2G This is a schematic timing diagram illustrating exemplary gate voltage signals according to various embodiments of the present disclosure;
[0016] Figure 3 This is a schematic block diagram illustrating exemplary signal conditioning circuits according to various embodiments of the present disclosure;
[0017] Figure 4 The illustration shows exemplary voltage relationships and stress zone curves associated with aging mechanisms according to various embodiments of this disclosure;
[0018] Figure 5A and Figure 5B This is a schematic diagram illustrating exemplary duty cycle modulation levels and their corresponding timing behaviors according to various embodiments of the present disclosure;
[0019] Figure 6A and Figure 6B This is a schematic diagram illustrating another exemplary duty cycle modulation level and its corresponding timing behavior according to various embodiments of the present disclosure;
[0020] Figure 7 This is a schematic circuit diagram illustrating exemplary programmable delay structures according to various embodiments of the present disclosure;
[0021] Figure 8 This is a schematic circuit diagram illustrating another exemplary programmable delay structure according to various embodiments of the present disclosure; and
[0022] Figure 9 The flowchart illustrates an exemplary operation of a method for manufacturing an apparatus according to various embodiments of the present disclosure, the apparatus including a DUT and a reliability stress circuit for aging assessment.
[0023] [Symbol Explanation]
[0024] 100: Exemplary device
[0025] 110: Device Under Test (DUT)
[0026] 120: Reliability stress circuit
[0027] 130: Frequency Generator
[0028] 140: Frequency divider
[0029] 150: Signal conditioning circuit
[0030] 160: Load circuit
[0031] 200: Exemplary device
[0032] 210: Device Under Test (DUT)
[0033] 220: Reliability stress circuit
[0034] 230: Frequency Generator
[0035] 240: Frequency divider
[0036] 250: Signal conditioning circuit
[0037] 260: Load circuit
[0038] 270: Frequency divider chain
[0039] 280: Switch
[0040] 290: Oscilloscope
[0041] 290' : Oscilloscope
[0042] 300: Indicative signal conditioning circuit
[0043] 310: Working Cycle Modulation Level
[0044] 320: Waveform Shaping Grade
[0045] 330: Signal Edge Shift Regulator
[0046] 340: Level shifter
[0047] 500: Working cycle modulation level
[0048] 600: Working cycle modulation level
[0049] 700: Programmable Delay Structure
[0050] 710: Selectable capacitor bank
[0051] 720: Encoder circuit
[0052] 800: Instance Delay Control Structure
[0053] 810: Encoder circuit
[0054] 820: Encoder circuit
[0055] 900: Method
[0056] 910: Operation
[0057] 920: Operation
[0058] 930: Operation
[0059] 940: Operation
[0060] 950: Operation
[0061] 960: Operation
[0062] 970: Operation
[0063] 980: Operation
[0064] A0, A1: Binary input
[0065] A: Node / Signal
[0066] B: Node / Signal
[0067] Cgd: Parasitic capacitance
[0068] Cgs: Parasitic capacitance
[0069] C: Signal
[0070] D0~D4: Control signals
[0071] D: Signal
[0072] f / 2: Output signal
[0073] f / 4: Output signal
[0074] f / 64: Output signal
[0075] f / n: Lower frequency output signal
[0076] f: High-frequency signal
[0077] HV: High drain voltage
[0078] Q: Signal
[0079] Ron: On-resistance
[0080] Vds: Drain-to-source voltage
[0081] Vdd: Voltage
[0082] VD: Drain voltage
[0083] VG1: Voltage
[0084] VG2: Voltage
[0085] Vgs: Gate-to-source voltage
[0086] VG: Gate voltage / gate drive signal
[0087] Vss: Grounding
[0088] Vt: Threshold voltage Detailed Implementation
[0089] The following disclosure provides numerous different embodiments or instances for implementing various features of the provided subject matter. Specific examples of components and configurations are described below to simplify this disclosure. Of course, these components and configurations are merely examples and are not intended to be limiting. For example, in the following description, the formation of a first feature over or on a second feature may include embodiments where the first and second features are formed in direct contact, and may also include embodiments where additional features may be formed between the first and second features such that the first and second features are not in direct contact. Furthermore, references to numbers and / or letters may be repeated in various instances of this disclosure. This repetition is for simplicity and clarity and does not in itself indicate a relationship between the various embodiments and / or configurations discussed.
[0090] Aging tests can be used to assess the long-term reliability of semiconductor devices by evaluating how the device under test (DUT) degrades over time when subjected to electrical or thermal stress. Electrical parameters, such as threshold voltage, drive current, leakage circuitry, on-resistance, and shifts in timing delays, can reflect aging-induced degradation and adversely affect device performance or ultimately cause failure during system operation.
[0091] In one type of aging test setup, a device under test (DUT), such as a power transistor, is biased according to its electrical condition. A signal generator applies a pulsed voltage to the gate terminal, while a power supply delivers a constant voltage to either the drain or source terminal. The remaining terminals may be connected to a reference potential or ground. The DUT is then subjected to a stress condition for an extended period of time, after which the electrical characteristics of the DUT are measured to indicate degradation.
[0092] However, such settings apply a fixed voltage condition and do not capture the dynamic switching conditions experienced by the device in a real-world operating environment. In many applications, semiconductor devices can operate with time-varying signals and experience transient voltage and current conditions that are not accurately represented in static bias tests.
[0093] The systems and methods described in some embodiments herein provide configurable stress waveforms, dynamic operation, a wider range of voltage and current control, and allow for the integration of aging and stress testing circuitry directly on the wafer or within the device. These capabilities facilitate more accurate simulation of real-world operating environments. Exemplary circuit configurations and calibration techniques are described in further detail below.
[0094] Figure 1A This is a schematic block diagram illustrating an exemplary device 100 according to various embodiments of the present disclosure. Figure 1AAs illustrated, the exemplary device 100 includes a DUT 110 and a reliability stress circuit 120, which subjects the DUT 110 to electrical stress conditions to assess the long-term reliability of the DUT 110. In some embodiments, the DUT 110 includes power supply devices such as high-voltage metal-oxide-semiconductor (HV-MOS) transistors, laterally diffused MOS (LDMOS) transistors, gallium nitride (GaN) transistors, or silicon carbide (SiC) transistors. In other embodiments, the DUT 110 may include other types of semiconductor components, such as standard logic cells, memory cells, analog front-end circuitry, or mixed-signal devices.
[0095] The reliability stress circuit 120 includes functional blocks that work together to generate and apply programmable gate and drain signals (VG, VD) to the DUT 110, allowing different electrical stress conditions to be simulated. Through waveform shaping, timing control, and coordinated voltage or current loads, the reliability stress circuit 120 supports applications that simulate stress profiles of real-world operating environments. This architecture enables repeatable and controlled aging assessments based on mechanisms such as hot-carrier injection (HCI), bias-temperature instability (BTI), and non-conductive stress (NCS). Further details of each block and its operation are described below.
[0096] In some configurations, device 100 may be formed on a semiconductor wafer comprising multiple integrated circuits (ICs) (or semiconductor dies), each IC separated by scribe lines (also referred to as "sawtooth streets" or "cut tracks"). Scribe lines are narrow, non-functional areas that facilitate wafer monolithization and may include alignment marks, test pads, or dummy features. In some embodiments, one or both of DUT 110 and reliability stress circuitry 120 are formed within a scribe line, thereby enabling stress testing without occupying functional die regions of the IC.
[0097] In other configurations, device 100 is implemented on a diced semiconductor die, wherein the DUT 110 and reliability stress circuitry 120 are fabricated in the active region of the IC and retained after dicing. This configuration allows stress testing to be performed after packaging and supports monitoring or post-deployment diagnostics within the system. While the scribing method benefits from minimal area impact and parallel wafer level testing, embedding the test circuitry in the active die provides enhanced flexibility in built-in self-test (BIST) and long-term reliability tracking.
[0098] Figure 1B This is a schematic block diagram illustrating an exemplary reliability stress circuit 120 according to various embodiments of the present disclosure. Figure 1B As illustrated, the example reliability stress circuit 120 includes a frequency generator 130, a frequency divider 140, a signal conditioning circuit 150, and a load circuit 160. These blocks cooperate to generate and apply a gate drive signal (VG) to the DUT 110 to simulate different electrical stress profiles. For example, the frequency generator 130 is implemented as a ring oscillator that outputs a high-frequency signal (f) used as a timing reference. The frequency divider 140 receives this high-frequency signal (f) and generates a plurality of lower-frequency output signals (f / n) (e.g., f / 2, f / 4, ..., f / 64), each lower-frequency output signal corresponding to a different switching condition (e.g., light, medium, heavy switching condition). One of the lower-frequency output signals (f / n) is routed to the signal conditioning circuit 150.
[0099] Signal conditioning circuit 150 processes the input frequency signal (f / n) and generates a gate drive waveform (VG) for DUT 110. As will be described in further detail below, this signal conditioning circuit 150 modifies waveform characteristics such as duty cycles, edge transitions, and output amplitude to mimic various switching conditions and stress scenarios.
[0100] Load circuit 160 is connected to the current-carrying terminals of DUT 110, such as the drain terminal, and applies controlled electrical stress during testing. Load circuit 160 operates in conjunction with the gate signal (VG) to supply current when DUT 110 is powered on. Current is effectively interrupted when the gate voltage is removed or reduced below a threshold, thus supporting time-based stress control. Depending on the test objectives, load circuit 160 also allows adjustment of stress levels, such as magnitude or duration. Load circuit 160 enables repeatable controlled stress for different DUTs and test scenarios.
[0101] Additionally, the load circuit 160 can mimic real-world parasitic effects, including trace inductance, wire resistance, and package capacitance. Combined with waveform control from the signal conditioning circuit 150, this configuration supports realistic stress profiles and facilitates testing of mechanisms such as HCI, BTI, and NCS. In some embodiments, the DUT 110 can be a power metal-oxide-semiconductor (MOS) device, such as an HV-MOS, LDMOS, GaN, or SiC transistor. For example, HCI stress is applied by setting the drain voltage high and the gate voltage to approximately half that of the drain voltage, thereby promoting carrier implantation at the gate oxide. BTI is induced by holding the gate terminal at a high (or negative) voltage while grounding the other terminals, thereby causing charge trapping. NCS is applied using a sub-critical gate voltage and a high drain bias to simulate shock ionization without strong conduction.
[0102] Figure 2A This is a schematic block diagram / circuit diagram illustrating another exemplary device 200 according to various embodiments of the present disclosure. Figure 2A As illustrated, example device 200 (e.g., device 100) includes a DUT 210 for stressing by a reliability stress circuit 220, which includes a frequency generator 230, a frequency divider 240, a signal conditioning circuit 250, and a load circuit 260. The frequency generator 230 generates a high-frequency output signal (f) (e.g., 640 MHz) that acts as a reference clock to generate timing signals used throughout the reliability stress circuit 220. In some embodiments, the frequency generator 230 may be implemented using a compact oscillator or a ring oscillator. Other types of frequency generators are contemplated in alternative implementations.
[0103] Frequency divider 240 receives a high-frequency output signal (f) from frequency generator 230 and generates multiple lower-frequency output signals (f / n) (e.g., f / 2, f / 4, ..., f / 64, such as 40 MHz, 20 MHz, and 10 MHz). These divided signals (f / n) can be selected to simulate different switching conditions, such as light, medium, or heavy switching activity. For example, frequency divider 240 may include a frequency division chain 270 formed from sequential logic elements (e.g., flip-flops, such as D-type flip-flops) and one or more switches 280 that selectively route the divided signals (f / n) to signal conditioning circuitry 250. Switches 280 may be implemented as gates, three-state buffers, or transmission gates, or alternatively, multiplexers, depending on design requirements, to simplify routing and enhance programmability.
[0104] The selected frequency division signal (f / n) is then processed by signal conditioning circuit 250, which generates a gate control signal (VG) applied to the gate of DUT 210. Signal conditioning circuit 250 can shape the input waveform and adjust parameters such as duty cycle, rate of change, and output voltage level. These modifications allow for the generation of a gate drive signal (VG) customized to mimic specific aging conditions and switching states. The output (VG) is then provided to the gate of DUT 210.
[0105] Load circuit 260 is connected to current-carrying terminals of DUT 210, such as the drain terminal, and is used to apply voltage or current stress according to the gate control signal (VG). In one embodiment, load circuit 260 includes two cascaded current mirrors and a current source. The first current mirror is connected to a high-voltage (HV) rail and mirrors a reference current downwards. A second current mirror, connected between the first current mirror and the current source, further replicates and modulates the current through DUT 210. This two-stage current mirror configuration improves current accuracy, thermal stability, and ensures symmetrical behavior, even under high-voltage operation.
[0106] The current source setting defines a reference current that determines the magnitude of the applied electrical stress. In some embodiments, the reference current ranges from approximately 0.1 mA to approximately 100 mA, depending on the required stress intensity and the characteristics of the DUT 210. The reference current can be adjusted according to the stress condition or test configuration. The load circuit 260 conducts when the DUT 210 is on (i.e., when VG is greater than the threshold voltage) and blocks conduction when the DUT 210 is off, thereby enabling a controlled pulse-based stress pattern and protecting the DUT 210 from excessive stress during non-operational intervals. To accommodate variations in manufacturing processes and temperature, the system allows tuning the switching frequency by adjusting the supply voltage. For example, when the drain voltage (VD) decreases to 0.96 V (-20%), the switching frequency is approximately 5.2 MHz; at a nominal VD of 1.2 V, the frequency is approximately 9.03 MHz; and at a high VD of 1.44 V (+20%), the frequency increases to approximately 12.3 MHz. This voltage-frequency control mechanism enhances the flexibility of maintaining constant stress applications regardless of PVT variations. Furthermore, the cascaded structure used in the load circuit enables safe operation at high voltages ranging from 1.8 V to 12 V, while also protecting the internal stress circuitry from excessive voltage fluctuations.
[0107] Through coordinated operation of signal conditioning circuit 250 and load circuit 260, device 200 can perform various electrical aging mechanisms. For example, hot-carrier injection (HCI) stress is induced by applying a high drain voltage (HV) via load circuit 260 while simultaneously setting the gate voltage (VG) to approximately half the drain voltage, which facilitates the injection of high-energy carriers into the gate oxide at the gate-drain overlap. Bias-temperature instability (BTI) is applied by maintaining a high gate voltage for an n-type device (or a negative gate voltage for a p-type device) while holding the source and drain terminals close to ground, resulting in charge trapping in the gate dielectric. Non-conductive stress (NCS) is achieved by applying a subcritical gate voltage (VG) (e.g., 155 mV or 300 mV) in conjunction with a high drain voltage (HV), thereby simulating stress effects such as trap generation or impact ionization without creating significant channel conduction. The mirror-based load circuit architecture provides safe HV operation and accurately replicates parasitic behaviors such as inductive ringing or capacitor delay, making it suitable for evaluating HV-MOS, LDMOS, GaN, or SiC power devices.
[0108] As described above, Device 200 enables applications with customized stress profiles across various device types and aging mechanisms. Adjustable parameters, such as waveform shape, switching frequency, duty cycle, and voltage amplitude, ensure a wide range of real-world stress conditions. External tools, such as oscilloscopes 290 and 290', can be connected to the gate and drain paths for real-time waveform observation and test verification.
[0109] Furthermore, the Reliability Stress Circuit 250 architecture uses a compact oscillator with a 1 / 64 divider to support high-frequency core operation. This configuration enables the generation of a wide range of stress frequencies while maintaining a small circuit footprint. The compact design allows the oscillator to be implemented within narrow wafer scribe lines between adjacent IC dies, thereby facilitating wafer-level reliability screening without occupying active die area. For example, in a 10 MHz oscillator design, an NMOS transistor can be configured with a length (L) of 0.6 μm, a width (W) of 2 μm, and a multiplier of 2, while a PMOS transistor can have the same length and width but with a reduced multiplier of 6. This optimization reduces the layout area and enables higher oscillator frequencies suitable for aging stress conditions, especially in space-constrained applications similar to wafer scribe lines.
[0110] Figure 2B This is a schematic timing diagram illustrating the exemplary relationship between frequency signals (f, f / n) according to various embodiments of the present disclosure. Figure 2BAs illustrated, the lower waveform represents the high-frequency output signal (f) operating at approximately 640 MHz, as generated by frequency generator 230. The upper waveform represents the lower-frequency output signal (f / 64) corresponding to 10 MHz in this example. The example diagram emphasizes how frequency divider 240 reduces the core oscillator signal to a slower, selectable switching frequency suitable for driving the DUT 210 under various stress conditions. The ability to scale down from a compact high-frequency oscillator to a precisely defined lower-frequency signal provides flexibility for customized test scenarios of different device types and reliability conditions.
[0111] Figure 2C This is a schematic timing diagram illustrating an exemplary lower-frequency output signal (f / n) generated from a high-frequency input signal (f) according to various embodiments of the present disclosure. Figure 2C The diagram shows waveforms representing three example outputs from divider 240: f / 16 = 40MHz (top), f / 32 = 20MHz (middle), and f / 64 = 10MHz (bottom). These lower frequency output signals (f / n) demonstrate how divider 240 scales down the output of frequency generator 230 to produce selectable switching frequencies suitable for different stress testing scenarios. By selecting these output frequencies, device 200 can mimic switching conditions with varying severity, such as light, moderate, or heavy device activity, thereby supporting flexible and programmable stress assessment of DUT 210.
[0112] Figure 2D This is a schematic timing diagram illustrating exemplary voltage level shifts performed by signal conditioning circuit 250 according to various embodiments of the present disclosure. Figure 2D As illustrated, the top waveform represents the input signal operating at approximately 1.2V, generated from the lower frequency output signal (f / n) received by the frequency divider 240. The bottom waveform represents the corresponding output gate control signal (VG), generated by the signal conditioning circuit 250 and shifted to a higher voltage level of 5V. This voltage shift allows the reliability stress circuit 220 to drive high-voltage devices, such as power transistors. The level shift can be achieved via a level shifter stage of the signal conditioning circuit 250 and can support various output levels (e.g., 1.8V, 3.3V, 5.0V, or sub-critical voltage levels such as 300mV or 155mV) depending on the aging stress profile of the application.
[0113] Figure 2E This is a schematic timing diagram illustrating exemplary gate signals (VG) according to various embodiments of the present disclosure. Figure 2EAs illustrated, the example gate signal (VG) has a 10% duty cycle, where the DUT 210 is turned on for only a short time interval during each cycle. This narrow pulse duration can be used to impose a simple periodic electrical stress representing the low operating, low conduction mode encountered during standby or idle operation. In some embodiments, the output (VG) is applied to the gate terminal of the DUT 210.
[0114] As described above, to adapt to process and temperature changes, devices 100 and 200 enable frequency tuning by adjusting the power supply voltage. For example, Figure 2F This is a timing diagram illustrating exemplary switching frequency variations in response to different drain voltages (VD) according to various embodiments of the present disclosure. Figure 2F As illustrated, when the drain voltage (VD) decreases to 0.96V (-20%), the frequency decreases to approximately 5.2MHz; at a nominal VD of 1.2V, the frequency is approximately 9.03MHz; and when VD increases to 1.44V (+20%), the frequency rises to approximately 12.3MHz. This voltage-frequency control scheme provides added flexibility in application-consistent stress conditions, even under varying PVT. By dynamically employing a switching frequency in response to the applied voltage, devices 100 and 200 help maintain a representative and controlled stress profile across changing operating environments and manufacturing conditions.
[0115] Figure 2G This is a schematic timing diagram illustrating exemplary gate voltage signals (VG) applied to DUT 210 according to various embodiments of the present disclosure. Figure 2G As illustrated, the signal conditioning circuit 250 can generate two unique instances of sub-critical gate voltage waveforms: one gate voltage waveform with a peak voltage of approximately VG1 = 155mV, and the other with a peak voltage of approximately VG2 = 300mV. These critical (VG) waveforms can be used to induce stress effects such as HCI and NCS. This allows aging effects to be tested under low gate voltage (VG) conditions.
[0116] Figure 3 This is a schematic block diagram illustrating an exemplary signal conditioning circuit 300 according to various embodiments of the present disclosure. The exemplary signal conditioning circuit 300 (e.g., signal conditioning circuit 250) receives a selected frequency-divided signal (e.g., f / 2, f / 4, f / 64) from a frequency divider 240 and generates a gate control signal (VG) applied to the gate of a DUT (e.g., DUT 110, 210). Figure 3As illustrated, the signal conditioning circuit 300 includes, in sequence, a duty cycle modulation stage 310, a waveform shaping stage 320, a signal edge transition adjuster 330, and a level shifter 340. These blocks work together to modify signal characteristics such as duty cycle, waveform shape, edge profile, and output amplitude, thereby enabling the application of gate signals simulating a wide range of electrical stress conditions.
[0117] The specific sequence of blocks 310 to 340 is illustrative and not limiting. In some implementations, the configuration may vary, certain functions may be omitted, or multiple functions may be combined depending on the test objective, available area, or characteristics of DUT 110, 210. The modular architecture provides flexibility for tuning across different test configurations and reusing design components.
[0118] The duty cycle modulation stage 310 adjusts the ratio between the high and low portions of each cycle of the lower frequency input signal (f / n) to mimic different switching activity levels. For example, 10% of the duty cycle provides a short gate-on pulse (mild stress), while 90% of the duty cycle provides a longer gate-on time interval (heavy stress). By varying the duty ratio, circuitry 310 controls how long the DUTs 110 and 210 remain on per cycle, directly affecting the energy delivered to the gate and the accumulated electrical stress. The waveform shaping stage 320 transforms, for example, the square wave of the duty modulation into a non-square waveform, such as a triangular, sawtooth, or trapezoidal waveform. These shapes mimic gate drive scenarios, including soft-switching or resonant gate drives, encountered in high-speed switching applications. For example, a trapezoidal or sawtooth gate signal can simulate slope-controlled switching behavior, thus better mimicking analog environments or EMI-sensitive systems.
[0119] Signal transition modulator 330 modifies the rising and falling edge rates of the shaped model to control the rate of change of the gate signal (VG). Fine-tuning the transition speed allows for the simulation of different aging conditions. Slower transitions emphasize BTI-related stresses, while faster transitions increase susceptibility to HCI. This block enhancement device 100, 200 tests the degradation mechanism's capability based on switching dynamic parameters. Level shifter 340 converts the shaped and transition control signals from an internal control voltage level (e.g., 1.2 V) to a higher or lower voltage domain suitable for gate drive.
[0120] This enables compatibility with a wide range of DUTs 110 and 210. For example, the level shifter outputs 3.3 V or 5.0 V, or lower voltages such as 300 mV or 155 mV, for HV-MOS or GaN devices to simulate non-conductive stress (NCS) or partial turn-on under HCI conditions. By enabling dynamic gate voltage scaling, this block ensures accurate reproduction of gate stress conditions relative to reliability testing.
[0121] Figure 4 The illustration shows exemplary voltage relationships and stress zone curves associated with aging mechanisms (HCI, BTI, NCS) according to various embodiments of this disclosure. Figure 4 As illustrated in the diagram, the vertical axis represents the gate-to-source voltage (Vgs), and the horizontal axis represents the drain-to-source voltage (Vds). The depicted curves define the operating patterns of the DUT (e.g., DUT 110, 210) under dynamic switching conditions. The upper curve reflects the actual switching response of the power device during circuit operation, taking into account intrinsic device behavior as well as parasitic capacitances (e.g., Cgd, Cgs) and degradation effects. The lower curve represents the ideal switching pattern assuming a purely capacitive response without parasitic effects.
[0122] The area marked "Capacitor" highlights the deviation between the actual switching path and the ideal switching path. This area can be interpreted as a voltage-time region, reflecting the effects of parasitic capacitance and switching delay. This area can be associated with energy distortion, charge hysteresis, or increased stress exposure during high-frequency or resonant operation. In detail, this deviation provides insight into the interaction between the device's intrinsic gate capacitance and external parasitic loads, both of which contribute to reliability degradation over time.
[0123] In this exemplary embodiment, reliability stress circuitry (e.g., circuits 100, 200) is used to generate waveforms (e.g., LOCUS) that allow for controlled lateral shifting of the Vgs-Vds plane. These waveforms mimic ideal switching conditions and allow for selective application of aging stress mechanisms. For example, HCI can be induced by applying a high Vds in conjunction with an intermediate Vgs. BTI can be triggered using a sustained high or negative gate bias, while NCS can be applied using a threshold gate voltage in conjunction with a high drain bias.
[0124] By fully utilizing the LOCUS waveform and adjustable bias control, devices 100 and 200 can simulate stress conditions and track parameters such as threshold voltage (Vt), on-resistance (Ron), and / or leakage current operational drift. The region between the ideal switching curve and the actual switching curve can be measured and quantified to calculate the figure of merit (FOM), which indicates the severity of aging or performance degradation. This allows reliability stress circuits 100 and 200 to facilitate aging assessment by simulating the stress degradation levels of DUTs 110 and 210 over time.
[0125] Figure 5A and Figure 5B This is a schematic diagram illustrating exemplary duty cycle modulation levels and their corresponding timing behaviors according to various embodiments of the present disclosure. For example... Figure 5A As illustrated, the example duty cycle modulation stage 500 receives the input signal at node (A). Stage 500 includes an even number of inverter stages that generate a delayed version of the input signal (A) at node (B). Because an even number of inverters are used, the signal at node (B) maintains the same logic level as signal (A). For example, when A is high, B is also high, although delayed in time due to the propagation delay through the inverter chain. This delay defines the programmable window used for duty cycle control.
[0126] Signal (A) is the original input signal, while signal B is the delayed version. The duty cycle percentage is determined based on the falling edge of A and the relative timing of B. Signal (C) represents the interval between the rising edge of B and the falling edge of A, during which A is low and B remains high. Signal (D) is the inverted version of signal (C), that is, D is high when C is low and low when C is high, and can be used as an enable signal to gate the output pulse.
[0127] In this configuration, signal (D) facilitates defining the actual duty cycle by determining the time it takes for a portion of the input signal (such as the interval between the falling edge of A and the delayed edge of B) to be allowed to propagate to the output. The combination of signals (A, B, C, D) enables flexible waveform cutting with fine-grained control, allowing for a programmable duty cycle ranging from approximately 2% to 98%. This tunable modulation architecture supports various electrical stress conditions for power device reliability assessments, including scenarios such as hot carrier injection (HCI) and non-conductive stress (NCS), by using gate signals to mimic switching activity.
[0128] Figure 6A and Figure 6BThis is a schematic diagram illustrating another exemplary duty cycle modulation level and its corresponding timing behavior according to various embodiments of the present disclosure. For example... Figure 6A As illustrated, the example working cycle modulation stage 600 receives the input signal at node (A) and includes an odd number of inverter stages that generate a delayed and inverted signal at node B. Because an odd number of inverters are used, signal B has the opposite logical polarity to signal (A), that is, when A is high, B is low, and vice versa, where the propagation delay is introduced by the inverter chain.
[0129] By adjusting the number of stages or the propagation delay between A and B, the modulation stage 600 supports flexible timing control, allowing the generation of gate pulses in programmable duty cycles. This configuration enables realistic operating stress conditions such as 2%, 10%, or up to 98% power device on-time duty cycles, thereby simulating light, medium, or heavy load conditions found in actual switching environments.
[0130] Signal (A) is the original input waveform, while signal (B) is delayed and is an inverted version of signal (A). Signal (C) represents the interval between the falling edge of B and the rising edge of A, corresponding to a mismatch window during which A is high while B is still low. Signal (D) is an inverted version of C and can be used as a gate signal to define the output pulse. In this configuration, signal (D) facilitates defining the duty cycle by determining the time it takes for the input waveform to propagate to the output. The duration of D, and therefore the duty cycle, can be adjusted by controlling the delay between A and B. This configuration provides control over switching stress levels for power device reliability testing by enabling the simulation of dynamic load conditions.
[0131] Figure 7 This is a schematic circuit diagram illustrating exemplary programmable delay structures according to various embodiments of the present disclosure. In at least one embodiment, the programmable delay structure 700 is adapted to the delay generation path of at least one of the duty cycle modulation stages 500, 600, wherein the delay between the input signal (A) and the delayed signal (B) is used to control the duty cycle. In each stage 500, 600, the programmable delay structure 700 adjusts the signal transition time at the output (node 1) of the first inverter in the series to tune the pulse width of the gate-controlled output signal (D).
[0132] like Figure 7As illustrated, the example delay control structure 700 includes a selectable capacitor bank 710 with values of 1C, 2C, 4C, and 16C, each capacitor connected between node 1 and ground (VSS) via a corresponding selector switch. Depending on which control wirings (D0, D1, D3, and D4) are confirmed, the capacitor bank 710 introduces a variable capacitive load at node 1. Increasing the total capacitance at node 1 results in a slower voltage transition via the inverter chain, thereby increasing the delay between signals (A, B) and modifying the duration of the duty cycle window.
[0133] Control signals (D0 to D4) are generated by encoder circuit 720. Encoder circuit 720 receives binary inputs (e.g., A0, A1) and outputs control signals (D0 to D4) to control capacitor bank 710. In some embodiments, encoder circuit 720 may include inverters and a network of logic gates to decode the binary inputs (A0, A1) for selecting patterns for discrete capacitors, thereby allowing digital programmability of the delay. This architecture supports both coarse and fine delay resolution by selectively combining different capacitor values.
[0134] By integrating a programmable delay structure with duty cycle modulation stages 500 and 600, the system can generate gate drive signals with an on-cycle range of approximately 2% to 98%. This supports dynamic operating conditions, such as the simulation of light, intermediate, or heavy switching loads. Therefore, the delay-tuned structure enhances the flexibility and accuracy of stress testing and aging assessment for semiconductor power devices, including HV-MOS, LDMOS, GaN, and SiC transistors.
[0135] Figure 8 This is a schematic circuit diagram illustrating another exemplary programmable delay structure according to various embodiments of the present disclosure. In at least one embodiment, the programmable delay structure 800 is adapted to the delay generation path of at least one of the analog stages 500, 600 of the duty cycle, wherein the delay between the input signal (A) and the delayed signal (B) is used to control the duty cycle. In each stage 500, 600, the programmable delay structure 800 adjusts the signal transition time at the output (node 1) of the first inverter in the series to tune the pulse width of the gate-controlled output signal (D).
[0136] like Figure 8As illustrated, the example delay control structure 800 includes a pair of encoder circuits 810 and 820 and a pair of corresponding selection switches. Each set of switches is connected to node 1 via transistor-based pull-down paths, which act as discharge channels with controllable resistance. The encoder circuits 810 and 820 independently generate control signals to activate one or more pull-down paths. When a particular switch is enabled, the selected transistor reduces the discharge resistance at node 1, thereby accelerating the voltage transition and reducing the delay between signals (A, B).
[0137] Encoder circuits 810 and 820 can each receive binary inputs (e.g., A0, A1) and generate control signals for their corresponding switching networks. These encoder circuits 810 and 820 may include inverters and logic gate networks that decode the binary inputs into discrete switching control patterns, thereby enabling programmable delay adjustment. By enabling different pull-down combinations, the effective timing of the signal transition at node 1 can be modulated with both coarse and fine resolution.
[0138] By integrating the programmable delay structure 800 with duty cycle modulation stages 500 and 600, the system can generate gate drive signals, with the duty cycle on ranging from approximately 2% to 98%. This supports dynamic switching conditions, such as simulating light, medium, or heavy load scenarios. As an alternative to the capacitor-based delay tuning 700, the transistor-based architecture 800 provides additional design flexibility and compactness for stress testing and aging assessment of semiconductor power devices, including HV-MOS, LDMOS, GaN, and SiC transistors.
[0139] Figure 9 The flowchart illustrates an exemplary operation of a method 900 for manufacturing an apparatus (e.g., apparatus 100 to 300) according to various embodiments of the present disclosure, the apparatus including a device under test (e.g., DUT 110, 210) and reliability stress circuits (e.g., reliability stress circuits 120, 220) for aging assessment. Further reference will now be made for ease of understanding. Figures 1A to 8 Describe instance method 900. It should be understood that method 900 applies to all methods except... Figures 1A to 8 Structures outside of those structures in the method. Additionally, it should be understood that additional operations may be provided before, during, and after method 900, and in alternative embodiments of method 900, some of the operations described below may be replaced or eliminated.
[0140] In operation 910, multiple integrated circuits (ICs) are fabricated in the functional area of the semiconductor wafer, each IC being marked by a scribe line. In operation 920, at least one of the scribe lines and the functional area, a device under test (DUT) (e.g., DUT 110, 210) and a reliability stress circuit (e.g., reliability stress circuit 120, 220) are also formed. These structures can be fabricated using standard semiconductor process flows suitable for advanced technology nodes.
[0141] In operation 930, a frequency generator (e.g., frequency generator 130, 230) of the reliability stress circuit generates a high-frequency signal (e.g., hundreds of MHz, such as 640 MHz), which serves as a reference clock for generating various timing signals used within the reliability stress circuit. The frequency generator can be implemented using a compact oscillator, such as a ring oscillator, to enable integration within tight layout constraints similar to wafer scribing lines.
[0142] In operation 940, the frequency divider of the reliability stress circuit (e.g., frequency dividers 140, 240) receives a high-frequency signal and generates multiple lower-frequency signals (f / n) (such as f / 2, f / 4, ..., f / 64). These outputs simulate different switching conditions (e.g., light, medium, or heavy switching). The frequency divider may include a flip-flop and a chain of divisions of one or more switches or multiplexers, which routes a selected one of the divided signals to a signal conditioning circuit (e.g., signal conditioning circuits 150, 250).
[0143] In operation 950, the selected frequency division signal is received by a signal conditioning circuit, which generates a gate control signal (VG) by passing the received signal through multiple stages, such as duty cycle modulation, waveform shaping, signal edge transition adjustment, and level shifting processing. These stages modify the timing, shape, transition slope, and output amplitude of the signal to produce a gate waveform that represents the real-world switching conditions and aging profile.
[0144] In operation 960, a gate control signal (VG) is applied to the gate terminal of the DUT, while load circuits (e.g., load circuits 160, 260) connected to the current-carrying terminals (e.g., drain) of the DUT apply voltage or current stress synchronously with the gate signal. The load circuits may include a cascaded current mirror structure for high-voltage isolation and accurate current replication, and the current source can be set to a reference stress current (e.g., approximately 0.1 mA to approximately 100 mA). The cascaded structure can handle operating voltages from 1.8V to 12V, thereby ensuring safe operation even under high-voltage stress conditions.
[0145] To account for process and temperature variations, the switching frequency can be tuned proportionally via voltage scaling. For example, at -20% Vdd (VD = 0.96V), the switching frequency can be approximately 5.2MHz; at nominal VD = 1.2V, the switching frequency can be approximately 9.03MHz; and at +20% Vdd (VD = 1.44V), the switching frequency can be approximately 12.3MHz. This adaptive mechanism ensures consistent stress application regardless of environmental or process fluctuations.
[0146] Depending on the applied stress conditions, various aging mechanisms can be selectively triggered in the DUT. For example, HCI can be induced by applying a high drain voltage and a gate voltage approximately half that of the drain voltage. BTI can be induced by maintaining a high (or negative) gate voltage while grounding other terminals. NCS can be applied by combining a high drain bias with a subcritical gate voltage (e.g., 155mV or 300mV) to simulate shock ionization without actually turning on the device.
[0147] In Operation 970, waveform signals at the gate and drain terminals of the DUT can be monitored using external test equipment (e.g., an oscilloscope) or embedded monitoring points. These observations can be used to verify waveform integrity, assess stress effects, and monitor age-related changes in parameters such as threshold voltage, drive current, and leakage current.
[0148] Finally, in Operation 980, the semiconductor wafer is diced along the scribe lines to separate the individual ICs. The DUT and the reliability stress circuitry embedded in the scribe lines can be removed during dicing. The DUT and the stress circuitry embedded in the active die region are retained after dicing for further evaluation after packaging, such as in-system reliability checks, built-in self-test (BIST), or field diagnostics.
[0149] In one embodiment, an apparatus includes a semiconductor wafer, a device under test (DUT), and a reliability stress circuit. The semiconductor wafer includes a plurality of integrated circuits and a plurality of scribe lines formed in a plurality of operational regions of the wafer, each integrated circuit being separated by a separate scribe line. The DUT is formed in at least one of the plurality of scribe lines. The reliability stress circuit, formed in at least one of the plurality of scribe lines, applies a stress signal to the DUT for aging assessment and includes a frequency generator, a frequency divider, a signal conditioning circuit, and a load circuit. The frequency generator generates a signal. The frequency divider is coupled to the frequency generator and generates a plurality of divided frequency signals. The signal conditioning circuit receives one of the plurality of divided frequency signals, selects one, and generates a gate control signal for application to a first terminal of the DUT. The load circuit is coupled to a second terminal of the DUT and applies a voltage or current stress according to the gate control signal.
[0150] In another embodiment, an apparatus includes a device under test (DUT) and a reliability stress circuit. The DUT is formed in an active region of an integrated circuit (IC). The reliability stress circuit is formed in the active region and applies a stress signal to the DUT for aging assessment, and includes a frequency generator, a frequency divider, a signal conditioning circuit, and a load circuit. The frequency generator generates a signal. The frequency divider is coupled to the frequency generator and generates a plurality of lower frequency signals. The signal conditioning circuit receives one of the plurality of lower frequency signals, selects one, and generates a gate control signal for application to a first terminal of the DUT. The load circuit is coupled to the current-carrying terminal of the DUT and applies a voltage or current stress according to the gate control signal.
[0151] In some embodiments, the signal conditioning circuitry includes: a duty cycle modulation circuit for adjusting the duration ratio between high and low states of an input signal to mimic different switching patterns for stress applications; a waveform shaping circuit for transforming the duty cycle modulation signal from one waveform shape to another to simulate analog-like switching behavior; a signal transition conditioning circuit for modifying the rising edge transition rate and falling edge transition rate of the waveform to control the signal change rate and replicate dynamic switching conditions; and a level shifter for transitioning a logic level waveform to a higher or lower voltage domain. In some embodiments, the frequency divider includes a frequency divider chain, including: sequential logic elements; and one or more switches for selectively routing one of the frequency divider signals to logic circuitry. In some embodiments, the frequency divider includes a frequency divider chain, including: sequential logic elements for generating signals with gradually decreasing frequencies; and a multiplexer for receiving outputs from the sequential logic elements and selectively routing one of the frequency divider signals to the signal conditioning circuitry in response to a control signal indicating a selected stress configuration. In some embodiments, the load circuit generates current and suppresses current generation when the device under test (DUT) is turned off, and the current flows through the DUT when the DUT is activated in response to a gate control signal. In some embodiments, the load circuit includes a cascaded current mirror structure, comprising: a lower transistor for receiving a reference current and mirroring the reference current through the DUT, and an upper transistor stacked above the lower transistor and connected to a voltage source, the upper transistor being biased to operate in saturation to isolate the lower transistor from high voltage stress conditions.
[0152] In another embodiment, a method for evaluating a device under test (DUT) formed in an active region of an integrated circuit includes the steps of: generating a first signal using a frequency generator; generating a plurality of lower frequency signals using a frequency divider coupled to the frequency generator; selecting one of the lower frequency signals and generating a gate control signal using a signal conditioning circuit; applying the gate control signal to the gate terminal of the DUT; and applying voltage or current stress to a current carrying terminal of the DUT using a load circuit, wherein the voltage or current stress is applied according to the gate control signal.
[0153] In some embodiments, the step of generating a gated signal by the signal conditioning circuit includes the following steps: adjusting the duration ratio between the high and low states of a selected lower frequency signal to mimic different switching patterns used in stress applications; transforming the modulated signal from one waveform shape to another to simulate analog switching behavior; modifying the rising edge transition rate and falling edge transition rate of the waveform to control the signal change rate and replicate multiple dynamic switching conditions; and using a level shifter to convert the waveform to a higher or lower voltage domain. In some embodiments, the step of generating a lower frequency signal by a frequency divider includes the following steps: using sequential logic elements in a frequency divider chain to generate a signal with a gradually decreasing frequency; and using one or more switches to selectively route one of the frequency divider signals. In some embodiments, the step of generating a lower frequency signal by a frequency divider includes the following steps: using sequential logic elements to generate a signal with a gradually decreasing frequency; and using a multiplexer to selectively route one of the frequency divider signals to the signal conditioning circuit, wherein the selected signal is determined based on a control signal indicating a selected stress configuration. In some embodiments, the step of applying voltage or current stress by the load circuit includes the following steps: generating a current flowing through the device under test (DUT) when the DUT responds to a gate control signal for activation; and suppressing the current when the DUT is turned off. In some embodiments, the step of applying voltage or current stress by the load circuit includes the following steps: mirroring a reference current through the DUT using a lower transistor of a cascaded current mirror structure; and biasing an upper transistor stacked above the lower transistor and connected to a voltage source to operate in saturation, thereby isolating the lower transistor from high voltage stress conditions.
[0154] In another embodiment, a method includes: forming a plurality of integrated circuits (ICs) in an operational region of a semiconductor wafer, each IC being defined by a separate scribe line; fabricating a reliability stress circuit and a device under test (DUT) coupled to the reliability stress circuit in at least one of the scribe lines; generating a first signal from the reliability stress circuit; dividing the signal into a plurality of lower frequency signals; selecting one of the plurality of lower frequency signals based on a test configuration; generating a gate control signal from the selected lower frequency signal; applying the gate control signal to a gate terminal of the DUT; and applying a voltage or current stress to a current carrying terminal of the DUT according to the gate control signal.
[0155] In some embodiments, the method further includes the steps of: fabricating a reliability stress circuit and a device under test (DUT) connected to the reliability stress circuit in at least one of the operating regions; dicing the semiconductor wafer along a scribe line; generating a first signal from the reliability stress circuit; dividing the first signal into lower frequency signals; selecting one of the lower frequency signals by logic circuitry based on a test configuration; generating a gate control signal from the selected divided signal by logic circuitry; applying the gate control signal to a first terminal of the DUT; and applying voltage or current stress to a second terminal of the DUT according to the gate control signal. In some embodiments, the method further includes the steps of: generating the lower frequency signal in a progressive manner; and routing the selected signal to a signal conditioning circuitry using a switch or multiplexer based on a control signal. In some embodiments, the step of applying voltage or current stress includes the steps of: replicating a reference current through the DUT using a cascaded current mirror structure, the cascaded current mirror structure including a lower transistor receiving the reference current and an upper transistor biased to saturate to isolate the lower transistor from high voltage stress. In some embodiments, the method further includes the step of: generating a gate control waveform with multiple variable duty cycles to apply mild, intermediate, or heavy switching stress during the aging assessment. In some embodiments, the method further includes the step of applying hot carrier injection stress by biasing the device under test (DUT) with a high drain voltage and a gate voltage approximately equal to half the drain voltage, thereby accelerating carrier injection into the gate oxide. In some embodiments, the method further includes the step of applying bias-temperature instability stress by maintaining the gate of the DUT at a high positive or negative voltage level while holding other terminals nearly grounded, thereby promoting charge trapping in the gate dielectric. In some embodiments, the method further includes the step of applying non-conduction stress by combining a high drain voltage with a primary threshold gate voltage to induce shock ionization or trap generation without achieving conduction through the DUT.
[0156] The foregoing outlines the features of several embodiments to enable those skilled in the art to better understand the nature of this disclosure. Those skilled in the art will understand that this disclosure can be readily used as a basis for designing or modifying other processes and structures for implementing the embodiments introduced herein and / or achieving the same objectives and / or advantages. Those skilled in the art will also recognize that such equivalent constructions do not depart from the spirit and scope of this disclosure, and that such equivalent constructions can be modified, substituted, and replaced herein without departing from the spirit and scope of this disclosure.
Claims
1. A semiconductor device, characterized in that, Include: A semiconductor wafer includes a plurality of integrated circuits formed in a plurality of functional regions of the semiconductor wafer and a plurality of scribe lines, each integrated circuit being separated by a separate scribe line; A device to be tested, which is formed in at least one of the plurality of scribe lines; and A reliability stress circuit, formed in at least one of the plurality of scribe lines and used to apply a stress signal to the device under test for evaluation, the reliability stress circuit comprising: A frequency generator, used to generate a first signal; A frequency divider, which is coupled to the frequency generator and is used to generate multiple frequency-divided signals; A signal conditioning circuit is used to receive one of the plurality of frequency-divided signals, select one, and generate a gate control signal for application to a first terminal of the device under test; and A load circuit is coupled to a second terminal of the device under test and is used to apply a voltage or current stress according to the gate control signal.
2. The semiconductor device as claimed in claim 1, characterized in that, The signal adjustment circuit includes: A duty cycle modulation circuit is used to adjust the duration ratio between the high and low states of the input signal to mimic different switching patterns used in stress applications. A waveform shaping circuit is used to transform the operating modulation signal from one waveform shape to another to simulate analog-like switching behavior. A signal conversion and adjustment circuit is used to modify the rising edge transition rate and falling edge transition rate of the waveform to control the signal change rate and replicate dynamic switching conditions; and A quasi-shifter is used to convert the logic level waveform to a higher or lower voltage domain.
3. The semiconductor device as claimed in claim 1, characterized in that, The frequency divider includes a frequency divider chain, which includes: Multiple sequential logic elements used to generate multiple signals with progressively decreasing frequencies; and A multiplexer for receiving multiple outputs from the plurality of sequential logic elements and selectively routing one of the plurality of frequency division signals to the signal conditioning circuit in response to a control signal indicating a selected stress configuration.
4. The semiconductor device as claimed in claim 1, characterized in that, The load circuit includes a first-stage current mirror structure, comprising: A lower transistor is used to receive a reference current and mirror that reference current through the device under test. An upper transistor is stacked on top of a lower transistor and connected to a voltage source. The upper transistor is biased to operate in a saturated state to isolate the lower transistor from multiple high-voltage stress conditions.
5. A method for evaluating a device under test formed in an active region of an integrated circuit, characterized in that, The method includes the following steps: Use a frequency generator to generate a first signal; A frequency divider coupled to the frequency generator is used to generate multiple lower frequency signals; Select one of the plurality of lower frequency signals and use a signal conditioning circuit to generate a gate control signal; The gate control signal is applied to a gate terminal of the device under test; and A voltage or current stress is applied to a current-carrying terminal of the device under test using a load circuit, wherein the voltage or current stress is applied according to the gate control signal.
6. The method as described in claim 5, characterized in that, The step of generating the gate control signal by the signal adjustment circuit includes the following steps: The duration ratio between the high and low states of the selected lower frequency signal is adjusted to mimic different switching patterns used in stress applications. The signal modulated by the operation is transformed from one waveform shape to another to simulate a switching behavior similar to that in an analog circuit. Modify the rising edge transition rate and falling edge transition rate of the waveform to control the signal change rate and replicate multiple dynamic switching states; and Use a one-bit quasi-shifter to convert the waveform into a higher or lower voltage domain.
7. The method as described in claim 5, characterized in that, The step of applying the voltage or current stress by the load circuit includes the following steps: A reference current is reflected through the device under test using a lower transistor mirror of a cascaded current mirror structure; and An upper transistor stacked above the lower transistor and connected to a voltage source is biased to operate in a saturated state, thereby isolating the lower transistor from multiple high-voltage stress conditions.
8. A method for evaluating a device under test formed in an active region of an integrated circuit, characterized in that, Includes the following steps: Multiple integrated circuits are formed in multiple functional regions of a semiconductor wafer, each integrated circuit being defined by a separate scribe line; A reliability stress circuit and a test device coupled to the reliability stress circuit are fabricated in at least one of the scribed lines. A first signal is generated by this reliability stress circuit; The first signal is divided into multiple lower frequency signals; One of the plurality of lower frequency signals is selected based on a test configuration; The institute selects a lower frequency signal to generate a gate control signal; The gate control signal is applied to a gate terminal of the device under test; and A voltage or current stress is applied to a current-carrying terminal of the device under test based on the gate control signal.
9. The method as described in claim 8, characterized in that, Further steps include: A reliability stress circuit and a device under test connected to the reliability stress circuit are manufactured in at least one of the plurality of operational regions. The semiconductor wafer is cut along the scribed lines; A first signal is generated by this reliability stress circuit; The first signal is divided into multiple lower frequency signals; One of the plurality of lower frequency signals is selected by a logic circuit based on a test configuration; The logic circuit generates a gate control signal from the selected frequency division signal; The gate control signal is applied to a first terminal of the device under test; and A voltage or current stress is applied to a second terminal of the device under test based on the gate control signal.
10. The method as described in claim 8, characterized in that, The step of applying the voltage or current stress includes the following steps: replicating a reference current through the device under test using a cascaded current mirror structure, the cascaded current mirror structure including a lower transistor that receives the reference current and an upper transistor that is biased and saturated to isolate the lower transistor from the high voltage stress.
