JPWO2016035388A1 - Semiconductor test apparatus and semiconductor test method - Google Patents

Abstract

When the temperature of the semiconductor element 91 is a variable, the data holding unit 5b holds correlation data with the inter-terminal voltage necessary for energizing the constant current I0, and the semiconductor element 91 is energized with the constant current I0 in the power cycle test. A temperature at which the temperature of the semiconductor element 91 is calculated based on the data collection unit 4 that collects the actual measurement data of the inter-terminal voltage at the time, the actual measurement data collected by the data collection unit 4, and the correlation data held by the data holding unit 5b And a calculation unit 5a.

Description

  The present invention relates to a semiconductor test apparatus and a semiconductor test method, and more particularly to a configuration for managing temperature in a power cycle test.

  The power cycle test is a device for evaluating the reliability around the electrode joint portion of the semiconductor element incorporated in the module (semiconductor device), by repeatedly applying a current to the semiconductor element and locally generating heat, The degree of deterioration around the electrode joint is evaluated. At that time, it is necessary to confirm that the temperature of the semiconductor element is within the assumed minimum temperature and maximum temperature, and it is desirable that the temperature of the semiconductor element can be accurately measured.

  Here, if the semiconductor element itself has a temperature sensor function, the temperature can be easily measured, but many of the semiconductor elements do not have a temperature sensor function. Moreover, since the surface of the semiconductor element in the semiconductor device is covered with the package material, the temperature cannot be measured using means such as an infrared camera.

  Therefore, a method has been proposed in which a minute current for temperature measurement is applied to the PN junction portion of the semiconductor element, and the temperature of the semiconductor element is estimated from the relationship between the voltage and temperature acquired in advance (for example, Patent Document 1). reference.).

JP 2000-111416 A (paragraphs 0012 to 0017, FIGS. 1 to 5)

  However, the method of estimating the temperature by energizing a minute current for temperature measurement requires a power supply for stably supplying the minute current and a circuit configuration for switching the minute current and the current for the power cycle in a timely manner. It becomes. On the other hand, the power cycle test is a test for evaluating reliability, and continuous operation is often performed for several days to several months. In this case, if the apparatus configuration becomes complicated, the operation of the test apparatus itself becomes unstable, and it becomes difficult to perform proper reliability evaluation.

  The present invention has been made to solve the above problems, and provides a semiconductor test apparatus and a semiconductor test method capable of performing a power cycle test by appropriate temperature management without complicating the apparatus configuration. The purpose is that.

  A semiconductor test apparatus according to the present invention is a semiconductor test apparatus for executing a power cycle test that periodically repeats energization and stop of a current on a semiconductor element mounted in the semiconductor device, wherein the temperature of the semiconductor element is set. A correlation data holding unit for holding correlation data with an applied voltage necessary for energization of the current when a variable is used; and an actual measurement of the applied voltage when the current is passed through the semiconductor element in the power cycle test A data collection unit for collecting data; and a temperature calculation unit for calculating the temperature of the semiconductor element based on the actual measurement data collected by the data collection unit and the correlation data held by the correlation data holding unit. It is characterized by that.

  A semiconductor test method according to the present invention is a semiconductor test method for executing a power cycle test that periodically repeats energization and stop of current on a semiconductor element mounted in a semiconductor device. A correlation data holding step for holding correlation data with an applied voltage necessary for energizing the current when temperature is a variable, and the applied voltage when the current is energized to the semiconductor element in the power cycle test A data collection step for collecting actual measurement data, a temperature calculation step for calculating a temperature of the semiconductor element based on the actual measurement data collected in the data collection step, and the correlation data held in the correlation data holding step; It is characterized by including.

  According to the present invention, since the element temperature can be calculated from the voltage applied to the semiconductor element for supplying the main current in the power cycle, it is possible to perform a power cycle test in which the temperature is properly controlled without complicating the device configuration. .

It is a figure which shows the structure of the semiconductor test apparatus concerning Embodiment 1 of this invention. It is a block diagram which shows the structure of the control part which comprises the semiconductor testing apparatus concerning Embodiment 1 of this invention. It is a figure which shows the time-dependent change of the energization current in a power cycle test, an applied voltage, and element temperature. It is a figure which shows the relationship between element temperature when a fixed electric current is sent through a semiconductor element, and the voltage between terminals. In the semiconductor testing apparatus and the semiconductor testing method concerning Embodiment 1 of this invention, it is a waveform diagram of the electric current and the voltage between terminals which shows the data acquisition timing immediately after the start of electric current conduction for calculating minimum temperature. In the semiconductor testing apparatus and the semiconductor testing method concerning Embodiment 2 of this invention, it is a wave form diagram of the electric current which shows the data acquisition timing at the time of an electric current stop, and the voltage between terminals for calculating the maximum temperature.

Embodiment 1 FIG.
1 to 5 are diagrams for explaining a semiconductor test apparatus and a semiconductor test method according to a first embodiment of the present invention. FIG. 1 illustrates a connection between a semiconductor test apparatus and a semiconductor device as a test body. FIG. 2 is a block diagram showing a configuration of a control unit which is a part of the semiconductor test apparatus. FIG. 3 is a graph showing temporal changes in energization current, applied voltage, and semiconductor element temperature (element temperature) that are repeated in the power cycle test performed by the test apparatus and the test method.

  FIG. 4 is a diagram showing the relationship between the terminal voltage and the element temperature at a predetermined gate voltage and energization current. FIG. 5 is a waveform diagram of current and terminal voltage showing data acquisition timing immediately after the start of current conduction for calculating the minimum temperature for each cycle in the power cycle test.

  First, before the detailed description of the semiconductor test apparatus and the semiconductor test method according to the first embodiment of the present invention, a power cycle test as a premise thereof will be described. Unlike the heat cycle test in which heating is performed from the outside, the power cycle test changes the temperature by causing the semiconductor element itself to generate heat by energizing the semiconductor element. For example, FIG. 3 shows a current value (upper stage), drain-source voltage (terminal voltage) when a predetermined current is periodically energized (repeated ON / OFF) to a MOSFET (Metal Oxide Semiconductor Field Effect Transistor). : Middle stage), and changes in the temperature of the MOSFET (element temperature: lower stage) over time.

  In the energization period Pa in which a current is applied, the element temperature increases due to heat generation of the element, and in the stop period Pr in which the energization is stopped, the element temperature decreases due to heat dissipation. Therefore, when the energization period Pa is switched to the stop period Pr, the maximum temperature TH of the period is reached, and when the stop period Pr is switched to the energization period Pa, the period becomes the minimum temperature TL. Such a cycle is repeated on the order of 100,000 times. However, for proper durability evaluation, it is required that the maximum temperature TH and the minimum temperature TL for each cycle be managed within a set range.

  As shown in FIG. 1, a semiconductor test apparatus 10 according to a first embodiment of the present invention includes a semiconductor device 90 (an electric circuit that includes a semiconductor element 91 (MOSFET) in the semiconductor device 90 and a power cycle test). The power source 1 is electrically connected to the drain terminal 91d and the source terminal 91s, and the gate power source 2 is connected to the gate terminal 91g and the source terminal 91s and applies the gate voltage. A voltmeter 3 for measuring an applied voltage (inter-terminal voltage) applied to the source terminal 91s and the drain terminal 91d, and a semiconductor device 90 (semiconductor element 91) such as an energization start signal or an energization stop signal output from the power supply 1. ) That collects data such as information indicating the energization state of the) and signals indicating the inter-terminal voltage output from the voltmeter 3 in association with time. The data collection unit 4, the overall operation of the power cycle test, the control unit 5 that calculates the above-described minimum temperature TL and maximum temperature TH based on the data from the data collection unit 4, the state of the power cycle test, etc. And a display unit 6 for displaying.

  The data collection unit 4 only needs to be capable of holding data associated with digitized time, such as a digital oscilloscope, for a certain period. For example, when a signal measured and output by the voltmeter 3 is input to the data collecting unit 4, the data collecting unit 4 outputs the voltage between the terminals and the time data when the signal is received to the control unit 5. To do.

  The control unit 5 is configured by installing software in a so-called microcomputer or the like, and as shown in FIG. 2, a data holding unit 5b that holds correlation data between a terminal voltage and an element temperature, which will be described later, and data collection Based on the data receiving unit 5c that receives data from the unit 4, the data received by the data receiving unit 5c, the correlation data held in the data holding unit 5b, and the energization state information, the element temperature and the minimum temperature are calculated. And a temperature calculation unit 5a.

  The control unit 5 may be provided with a part having a function of controlling the operation of the power supply 1 and the gate power supply 2 as the operation of the power cycle test itself. Therefore, the description thereof is omitted. It should be noted that whether or not the part for calculating the temperature and the part for controlling the experiment itself is integrated does not limit the present invention and may be configured in any manner.

  The correlation data between the element temperature and the inter-terminal voltage held in the data holding unit 5b uses the element temperature as a parameter, the gate voltage is constant, and the inter-terminal voltage data necessary for supplying a predetermined drain current is obtained in advance. Get by measuring. Specifically, the semiconductor device 90 to be subjected to the power cycle test is held at a constant temperature in a thermostatic bath or on a hot plate, and a predetermined current is supplied under a condition that a gate-source voltage (gate voltage) is constant. Apply the value pulse current and measure the drain-source voltage (terminal voltage). If the pulse energization time is long, the element temperature changes more greatly than the holding temperature during measurement due to the self-heating effect.

  Therefore, the pulse energization time is set so that the change (increase) from the constant temperature of the semiconductor element 91 in the semiconductor device 90 to be tested held in the thermostatic bath or on the hot plate is within an allowable range. Adjusted. That is, the correlation data between the terminal voltage and the element temperature is obtained using the temperature of the thermostatic chamber or the hot plate as a parameter, but the element temperature is obtained as a parameter by adjusting the pulse time described above.

  Thus, when the obtained gate voltage and the drain current are constant, the relationship between the terminal voltage and the element temperature is a one-to-one relationship as shown in FIG. For example, the element temperature is uniquely determined. Therefore, in the data holding unit 5b, correlation data between the terminal voltage and the element temperature in the drain current and the gate voltage corresponding to the power cycle test of the semiconductor device 90 to be tested is, for example, an LUT (Look Up Table). Is held in a shape like As a result, the temperature calculation unit 5a is operated at least during the energization period Pa based on the voltage data received by the data reception unit 5c and the correlation data held in the data holding unit 5b. Can be calculated.

  Since the maximum temperature TH is an element temperature at the time of switching from the energization period Pa to the stop period Pr, it can be easily calculated based on the last voltage data in the energization period Pa. On the other hand, since the minimum temperature TL is a temperature in a state where current is not supplied, from the data of the voltage between two or more points immediately after the start of current conduction when switching to the current supply period Pa (different in time), the following is performed. calculate.

  In principle, two or more temperatures obtained from two or more terminal voltage values immediately after the start of current conduction are measured with a quadratic function of the time at which each temperature is measured (the current conduction start time is 0). Then, the zeroth order value of the quadratic function can be set as the minimum temperature TL. When the minimum temperature TL is calculated in this way, the energization start signal of the power source 1 becomes a data collection trigger signal for calculating the minimum temperature TL. At this time, the delay time Δt is set for the trigger signal so that the data acquisition unit 4 acquires data at a timing when the current becomes a constant value (excluding an unstable period immediately after the start of conduction). ing.

Then, in the semiconductor test apparatus 10 or the semiconductor test method according to the first embodiment, considering the conditions described later, as shown in FIG. TL was calculated. In the figure, an energization start signal (trigger signal) from the power source 1 in a certain cycle is input to the data collection unit 4 at time t ₀ . Then, the voltage data V ₁ at the first point and the time data t ₁ that is data of the measurement time are acquired by the data collection unit 4 at a time after the delay time Δt from the time t ₀ . As described above, the delay time Δt is determined by the time during which the current becomes a constant value (constant current I ₀ ), which is a value specific to the thermal connection environment between the power supply 1 and the semiconductor device 90 (semiconductor element 91). It can be decided by evaluating in advance.

Then, data of the second point is collected at the time t ₂ after the time t _1. The interval t ₂ -t ₀ between the time t ₀ and the time t ₂ is desirably set to a time sufficiently shorter than the temperature change time constant τ = R × C of the semiconductor device 90 (semiconductor element 91). R is the transient thermal resistance of the semiconductor element 91, and C is the heat capacity around the semiconductor element 91. In a model in which the temperature (element temperature) of the semiconductor element 91 is Tm and the transient thermal resistance and heat capacity around the semiconductor element 91 are regarded as constants, the time change of Tm is described by the following equation (1).

Here, time t is a value when time t ₀ = 0 seconds. T ₀ is the element temperature at time t ₀ , and T _f is the saturation temperature. At this time, if the interval t ₂ -t ₀ is sufficiently shorter than the temperature change time constant τ, the equation (1) is a linear approximation equation represented by the following equation (2).

If the time change of Tm is approximated by the linear form shown in Equation (2), T ₀ can be obtained if there are two points of temperature data at different times. For example, if the semiconductor element 91 is a MOSFET, drain-source voltages V ₀ , V ₁ , V ₂ at times t ₀ , t ₁ , t ₂ are output to the control unit 5 (data receiving unit 5c) as terminal voltages. Is done. Since no current flows at time t ₀ , V ₀ = 0V, and it is not necessary to measure the drain-source voltage V ₀ at t ₀ .

Then, T ₀ is calculated from the voltage data V ₁ and V _{2 at} times t ₁ and t ₂ as follows. The control unit 5 (temperature calculation unit 5a) calculates element temperatures T ₁ and T ₂ at times t ₁ and t ₂ from the voltage data V ₁ and V ₂ and the correlation data held in the data holding unit 5b, respectively. Further, T ₀ is calculated from the time difference data t ₁ -t ₀ , t ₂ -t ₀ and the calculated element temperatures T ₁ , T ₂ using the equation (2). The calculated T ₀ becomes the minimum temperature TL of the semiconductor element 91 in the cycle. The calculated minimum temperature TL is displayed on the display unit 6. The calculated minimum temperature TL and maximum temperature TH may be stored separately as a history of the power cycle test.

In the above, the terminal between the second data acquisition time t ₂ and t ₀ is made shorter than the temperature change time constant τ, and the terminal is based on the equation (2) obtained by linear approximation of the equation (1). A method has been described in which the minimum temperature TL is calculated by a simple calculation by reducing the number of inter-voltage measurement points. However, the minimum temperature TL can be calculated with higher accuracy by using an equation obtained by approximating the equation (1) with a quadratic equation. However, in this case, the data collection unit 4 measures the inter-terminal voltage at least three points at different times, and the control unit 5 (temperature calculation unit 5a), based on the time at each of the three points and the inter-terminal voltage data, Calculate the minimum temperature TL.

As described above, by measuring the voltage between the terminals when the main current is supplied to the semiconductor element 91, such as the drain-source voltage of the MOSFET, at least two points with different times immediately after the start of current conduction, the minimum temperature TL can be reduced. Can be calculated. Further, the maximum temperature TH can be easily obtained by measuring the voltage at one point immediately before the current is stopped. That is, the semiconductor test apparatus 10 or the semiconductor test method capable of performing the managed power cycle test by calculating the maximum temperature TH and the minimum temperature TL in each cycle by a simple method is obtained. In addition, although the example which approximates element temperature to a time function was demonstrated in the said example, it does not restrict to this. For example, the element temperature may be calculated based on the inter-terminal voltage at time t ₀ calculated by approximating the inter-terminal voltage to a time function.

  In particular, in the method of estimating the element temperature by supplying a small current for temperature measurement to the PN junction as described in the background art, a power source and a small current for stably supplying a small current for temperature measurement An additional circuit is required to control this. However, the semiconductor test apparatus 10 according to the first embodiment of the present invention does not require a power supply for stably supplying a minute current and an additional circuit for controlling the minute current, and the apparatus configuration is complicated. There is nothing.

  The semiconductor element used in the semiconductor device to be tested may be a general element based on a silicon wafer. However, the operation temperature is higher when a so-called wide band gap semiconductor material having a wider band gap than silicon carbide (SiC), gallium nitride-based material (GaN), or diamond is used. Reliability becomes important. When the semiconductor test apparatus 10 or the test method according to the first embodiment of the present invention is used, the effect of appropriately evaluating the durability and the like and obtaining a highly reliable semiconductor device becomes remarkable. The type of semiconductor element may be a switching element such as the above-described MOSFET or IGBT (Insulated Gate Bipolar Transistor), or a rectifying element such as a diode. Further, when an IGBT is used as the semiconductor element 91, the names of the terminals used in this embodiment may be replaced as follows. Source (source terminal 91s) → emitter, drain (drain terminal 91d) → collector.

As described above, according to the semiconductor test apparatus 10 according to the present embodiment, the power cycle test is periodically performed on the semiconductor element 91 mounted in the semiconductor device 90 to repeatedly energize and stop the current I _0. A correlation data holding unit (data holding) that holds correlation data with an applied voltage (inter-terminal voltage) necessary for energizing the current I ₀ when the temperature of the semiconductor element 91 is a variable. Unit 5b), a data collection unit 4 that collects actual measurement data of the applied voltage (voltage between terminals) when the current I ₀ is supplied to the semiconductor element 91 in the power cycle test, and actual measurement data collected by the data collection unit 4 And a temperature calculation unit 5a for calculating the temperature of the semiconductor element 91 based on the correlation data held by the correlation data holding unit (data holding unit 5b). A power supply for stably supplying a current and an additional circuit for controlling a minute current are not required, and the device configuration is not complicated. A cycle test can be performed.

Further, when the data collection unit 4 receives an energization start signal for starting energization of the semiconductor element 91 in each cycle in the power cycle test, the data collection unit 4 starts from the reception time (time t ₀ ) after receiving the energization start signal. Two or more data sets (V ₁ , t ₁ ) and (V ₂ , t ₂ ) of the actual measurement data and the measurement time at the time when the elapsed times of the two are different are collected, and the temperature calculation unit 5a collects the two collected sets Since the temperature of the semiconductor element 91 at the reception time (time t ₀ ) calculated by approximating the above data set to a function of time is set to the minimum temperature TL in each cycle, the period from the stop period Pr to the energization period Pa is set. It is the temperature at the time of switching, and the minimum temperature TL that is important for determining the appropriateness of the power cycle test can be easily calculated.

  In particular, the elapsed time is set to a time shorter than the temperature change time constant τ of the semiconductor element 91 in the semiconductor device 90, and the temperature calculation unit 5a approximates two or more collected data sets to a linear function of time. For example, the minimum temperature TL can be calculated even with two data sets, and the minimum temperature TL can be calculated more easily.

In addition, according to the semiconductor test method according to the first embodiment, a semiconductor test for executing a power cycle test in which the current I ₀ is periodically turned on and off is performed on the semiconductor element 91 mounted in the semiconductor device 90. A method for holding correlation data with an applied voltage (inter-terminal voltage) necessary for energizing the current I ₀ when the temperature of the semiconductor element 91 (executed before the power cycle test) is a variable. A correlation data holding step (performed before the power cycle test) and a data collection step for collecting measured data of applied voltage (voltage between terminals) when the current I ₀ is applied to the semiconductor element 91 in the power cycle test. And a temperature calculation step of calculating the temperature of the semiconductor element 91 based on the actual measurement data collected in the data collection step and the correlation data held in the correlation data holding step. Thus, the power supply for stably supplying a minute current and an additional circuit for controlling the minute current are not necessary, and the temperature of the semiconductor element 91 can be increased without complicating the device configuration. The power cycle test can be performed with proper management.

Further, in the data collection process, when an energization start signal for starting energization of the semiconductor element 91 is received in each cycle in the power cycle test, from the reception time (time t ₀ ) after receiving the energization start signal, respectively. Collect two or more sets (V ₁ , t ₁ ) and (V ₂ , t ₂ ) of data sets (V ₁ , t ₁ ) and actual measurement data at the time points when the elapsed times are different. Since the temperature of the semiconductor element 91 at the reception time (time t ₀ ) calculated by approximating the data set to a function of time is set to the minimum temperature TL in each cycle, the switching from the stop period Pr to the energization period Pa is performed. The minimum temperature TL that is important for judging the suitability of the power cycle test can be easily calculated.

  In particular, the elapsed time is set to a time shorter than the temperature change time constant τ of the semiconductor element 91 in the semiconductor device 90, and in the temperature calculation step, two or more collected data sets are approximated to a linear function of time. For example, the minimum temperature TL can be calculated with two data sets, and the minimum temperature TL can be calculated more easily.

Embodiment 2. FIG.
In the first embodiment, only the description has been given of calculating the maximum temperature from the last voltage data in the energization period. In the second embodiment, the acquisition timing of the last voltage data is set according to the scheduled switching timing. FIG. 6 is a waveform diagram of the current and the voltage between the terminals for showing the data acquisition timing at the time of stopping the current for calculating the maximum temperature in the semiconductor test apparatus and the semiconductor test method according to the second embodiment of the present invention. is there. In addition, about another basic structure, it is the same as that of Embodiment 1, While using a figure, description is abbreviate | omitted about the same part.

In the figure, when switched to the stop period Pr from energization period Pa, the current stop trigger signal from the power supply 1 is inputted to the data collecting section 4 to the time t _x. In this case, ideally, the time t _x previously, holds the current constant value (constant current I _0), and to measure the terminal voltage at the time close to the time t _x desirable. However, for this, it is necessary to collect time data going back from the time t _x, for example, to measure the voltage data before the switching timing at a fine measurement period, is stored and read in FIFO (first-in, first-unloading) format It is necessary to provide a possible measurement / storage device.

On the other hand, unlike the case where switching to the conduction period Pa from stop period Pr, in the conventional power cycle test conditions, it is not the element temperature in the vicinity of the time t _x changes abruptly. In the present exemplary semiconductor test apparatus according to Embodiment 2, and semiconductor testing method, actually not the current stop trigger signal output to stop a current, based on the switching timing t _s which is planned as a power cycle test and to set the acquisition timing t _a to acquire voltage data for calculating the highest temperature TH. Acquisition timing t _a, even if the actual switching time t _x and planned switching timing t _s is shifted, without become later than t _x, and current is held at a constant value (a constant current I _0), And it is determined so as to be close to the time t _x .

Information of switching timing t _s is obtained, for example, may be held in advance data holding portion 5b and the like. For example, like the case of adjusting the timing for each cycle, the control unit for controlling the not-shown test You may make it do. With this configuration, the data collection unit 4, to the switching timing t _s which is scheduled in the period, to set the time of the predetermined period before the acquisition timing t _a, becomes the acquisition timing t _a, the terminal collect voltage data V _a between the voltage, and outputs it to the control unit 5.

In the control unit 5 (temperature calculation unit 5a), in a manner similar to that described in the first embodiment, based on the correlation data held in the voltage data V _a and the data holding portion 5b is input, at the acquisition timing t _a The element temperature is calculated. The calculated element temperature can be set as the maximum temperature TH in the cycle. For example, the calculated maximum temperature TH is displayed on the display unit 6. Also in the second embodiment, the calculated minimum temperature TL and maximum temperature TH may be separately stored as a power cycle test history.

That is, the data required terminal voltage to obtain a maximum temperature TH requires only one point, for example, there is no need to collect a large number of data that may correspond to the acquisition timing t _a. That is, it is not necessary to use a measuring instrument that requires high-speed storage capacity, and the maximum temperature TH for each cycle in the power cycle test can be managed using a simple measuring instrument. That is, by combining with the method for calculating the minimum temperature TL described in the first embodiment, a semiconductor test apparatus 10 or a semiconductor test method capable of performing a power cycle test under appropriate temperature management is obtained with a simpler apparatus. Can do.

As described above, according to the semiconductor test apparatus 10 according to the second embodiment, the data collection unit 4 has the scheduled stop time (switching timing t _s ) for stopping energization scheduled in each cycle in the power cycle test. based on the information, collects the measured data (voltage data V _a) at the scheduled stop time (switching timing t _s) from the predetermined time before the time (acquisition time t _a), the temperature calculation unit 5a, measured at the acquisition timing t _a Since the temperature of the semiconductor element 91 calculated based on the data (voltage data V _a ) is set to the maximum temperature TH in each cycle, it is the temperature when the energization period Pa is switched to the stop period Pr, and the power It is possible to easily calculate the maximum temperature TH that is important for determining the suitability of the cycle test.

Further, according to the semiconductor test method according to the present embodiment, in the data collection process, the scheduled stop based on the information on the scheduled stop time (switching timing t _s ) for stopping the energization scheduled in each cycle in the power cycle test. Actual measurement data (voltage data V _a ) at _a time (acquisition timing t _a ) _a predetermined time before the time (switching timing t _s ) is collected, and in the temperature calculation step, actual measurement data (voltage data V _a ) at the acquisition timing t _a is collected. Since the temperature of the semiconductor element 91 calculated based on the above is configured to be the maximum temperature TH in each cycle, the temperature at the time of switching from the energization period Pa to the stop period Pr is determined, and the appropriateness of the power cycle test is determined. It is possible to easily calculate the maximum temperature TH that is important for this.

1: power supply, 2: gate power supply, 3: voltmeter, 4: data collection unit, 5: control unit, 5a: temperature calculation unit, 5b: data holding unit, 5c: data reception unit, 6: display unit, 10: Semiconductor test equipment, 90: semiconductor equipment, 91: semiconductor element, 91d: drain terminal (power terminal), 91g: gate terminal (control terminal), 91s: source terminal (power terminal), I ₀ : constant current, Pa: energization period, Pr: stop period, _{t 0:} time (toggle its timing energizing period), _{t a:} obtaining timing, TH: reconversion temperature, TL: minimum temperature, time _{t 0:} time (reception _{time), t} s: (Scheduled) switching timing (scheduled stop time), t _x : (actual) switching timing, V ₁ , V ₂ : voltage data (measured data), V _a : voltage data (measured at the acquisition timing) Data), Δt: delay time.

A semiconductor test apparatus according to the present invention is a semiconductor test apparatus for executing a power cycle test that periodically repeats energization and stop of a current on a semiconductor element mounted in the semiconductor device, wherein the temperature of the semiconductor element is set. A correlation data holding unit for holding correlation data with an applied voltage necessary for energization of the current when a variable is used; and an actual measurement of the applied voltage when the current is passed through the semiconductor element in the power cycle test A data collection unit for collecting data; and a temperature calculation unit for calculating the temperature of the semiconductor element based on the actual measurement data collected by the data collection unit and the correlation data held by the correlation data holding unit. The data collection unit receives an energization start signal for starting energization of the semiconductor element in each cycle in the power cycle test. Collect two or more data sets of the actual measurement data and the measurement time at times when the elapsed time from the reception time after receiving the energization start signal is different, and the temperature calculation unit collects the collected 2 The temperature of the semiconductor element at the reception time calculated by approximating a data set equal to or more than a set to a function of time is set as a minimum temperature in each cycle .

A semiconductor test method according to the present invention is a semiconductor test method for executing a power cycle test that periodically repeats energization and stop of current on a semiconductor element mounted in a semiconductor device. A correlation data holding step for holding correlation data with an applied voltage necessary for energizing the current when temperature is a variable, and the applied voltage when the current is energized to the semiconductor element in the power cycle test A data collection step for collecting actual measurement data, a temperature calculation step for calculating a temperature of the semiconductor element based on the actual measurement data collected in the data collection step, and the correlation data held in the correlation data holding step; only contains, in the data collection process, in each cycle of the power cycle test, energized open to start energization of the semiconductor element When the signal is received, two or more data sets of the actual measurement data and the measurement time at the time when the elapsed time from the reception time is different after receiving the energization start signal are collected, and in the temperature calculation step, The temperature of the semiconductor element at the reception time calculated by approximating the collected two or more data sets to a function of time is set as the lowest temperature in each cycle .

Claims (8)

A semiconductor test apparatus that performs a power cycle test that periodically repeats energization and stop of current for a semiconductor element mounted in a semiconductor device,
A correlation data holding unit that holds correlation data with an applied voltage necessary for energization of the current when the temperature of the semiconductor element is a variable;
In the power cycle test, a data collection unit that collects actual measurement data of the applied voltage when the current is supplied to the semiconductor element;
A temperature calculation unit that calculates a temperature of the semiconductor element based on the actual measurement data collected by the data collection unit and the correlation data held by the correlation data holding unit;
A semiconductor test apparatus comprising: When the data collection unit receives an energization start signal for starting energization of the semiconductor element in each cycle in the power cycle test, the elapsed time from the reception time after receiving the energization start signal is different. Collect two or more data sets of the actual measurement data and the measurement time at the time,
The temperature calculation unit is characterized in that the temperature of the semiconductor element at the reception time calculated by approximating the collected two or more data sets to a function of time is the lowest temperature in each cycle. The semiconductor test apparatus according to 1. The elapsed time is set to a time shorter than the temperature change time constant of the semiconductor element in the semiconductor device,
The semiconductor test apparatus according to claim 2, wherein the temperature calculation unit approximates the collected two or more data sets to a linear function of time. The data collection unit collects the actual measurement data in a predetermined time before the scheduled stop time based on information on the scheduled stop time to stop the energization scheduled in each cycle in the power cycle test,
4. The temperature calculation unit according to claim 1, wherein the temperature of the semiconductor element calculated based on the actual measurement data at the time before the predetermined time is set as the maximum temperature in each cycle. The semiconductor test apparatus according to item 1. A semiconductor test method for performing a power cycle test that periodically repeats energization and stop of current for a semiconductor element mounted in a semiconductor device,
A correlation data holding step for holding correlation data with an applied voltage necessary for energization of the current when the temperature of the semiconductor element is a variable;
In the power cycle test, a data collection step of collecting measured data of the applied voltage when the current is supplied to the semiconductor element;
A temperature calculating step for calculating a temperature of the semiconductor element based on the measured data collected in the data collecting step and the correlation data held in the correlation data holding step;
A semiconductor test method comprising: In the data collection step, when an energization start signal for starting energization of the semiconductor element is received in each cycle in the power cycle test, the elapsed time from the reception time after receiving the energization start signal is different. Collect two or more data sets of the actual measurement data and the measurement time at the time,
The temperature calculation step is characterized in that the temperature of the semiconductor element at the reception time calculated by approximating the collected two or more data sets to a function of time is the lowest temperature in each cycle. 5. The semiconductor test method according to 5. The elapsed time is set to a time shorter than the temperature change time constant of the semiconductor element in the semiconductor device,
The semiconductor test method according to claim 6, wherein in the temperature calculation step, the two or more collected data sets are approximated to a linear function of time. In the data collection step, based on the information on the scheduled stop time for stopping the energization scheduled in each cycle in the power cycle test, the actual measurement data is collected at a predetermined time before the scheduled stop time,
8. The temperature calculation step according to claim 5, wherein the temperature of the semiconductor element calculated based on the actual measurement data at the time before the predetermined time is set as a maximum temperature in each cycle. 9. The semiconductor test method according to 1.

Images