Temperature estimation device for semiconductor devices

Abstract

The present invention provides a semiconductor device temperature estimation device capable of estimating the temperature of a semiconductor device in which the voltage drop exhibits the same value at multiple temperatures. [Solution] The semiconductor device temperature estimation device 100 comprises a heating power supply 1 that supplies a heating current Ih to the DUT 8, a measuring power supply 2 that supplies a measuring current Im to the DUT, and a control device 5 that acquires time-series data of the voltage drop of the first measurement target area of ​​the DUT. Prior to this, for the second semiconductor device to be measured, a temperature / voltage drop curve is measured in which the voltage drop of the second measurement target area changes with at least one of a maximum and minimum value as the temperature of the second measurement target area decreases. The control device 5 utilizes the fact that the temperature of all second measurement target areas in the temperature / voltage drop curve corresponds to the voltage drop of one second measurement target area, and converts the time-series data of the voltage drop of the first measurement target area of ​​the DUT into time-series data of the temperature of the said first measurement target area.

Description

[Technical Field]

[0001] This invention relates to a temperature estimation device for semiconductor devices. [Background technology]

[0002] Conventionally, because it is not possible to directly and accurately measure the junction temperature of a semiconductor device, a method for estimating the junction temperature of a semiconductor device is known, which involves measuring a physical quantity correlated with the junction temperature of the semiconductor device and converting the measured physical quantity into the junction temperature using the correlation between the physical quantity and the junction temperature. In a typical method for estimating the junction temperature of a semiconductor device, the voltage drop due to the measurement current at the junction of the semiconductor device is measured, and this measured voltage drop is converted into a temperature based on the temperature dependence of the voltage drop, and this temperature is estimated as the temperature of the junction (see, for example, Patent Document 1). [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2000-111416 [Overview of the Initiative] [Problems that the invention aims to solve]

[0004] Incidentally, in typical semiconductor device junction temperature estimation methods, the forward voltage drop due to the measurement current at the PN junction is measured. If the semiconductor device is a MOSFET, the forward voltage drop of the body diode, Vsd, is measured. However, the part of a MOSFET whose temperature we actually want to measure is the channel (inversion layer) between the drain and source, through which the load current flows during actual use. Therefore, it can be said that the method of measuring Vsd of a MOSFET cannot accurately estimate the temperature of the MOSFET. For this reason, methods have been proposed to estimate the temperature of a MOSFET, such as measuring the drain-source voltage Vds of the MOSFET and measuring the threshold voltage Vth (gate-source voltage Vgs) of the MOSFET. Furthermore, for IGBTs, which are devices that combine the MOSFET structure of the input section and the BIPOLAR structure of the output section, a method has been proposed to measure the threshold voltage Vth (gate-emitter voltage Vge).

[0005] On the other hand, in recent years, MOSFETs using SiC (silicon carbide), which have excellent performance (hereinafter abbreviated as "SiCMOSFET"), have been widely used. However, in SiCMOSFETs, the threshold voltage Vth (gate-source voltage Vgs) changes each time it is measured. Similarly, the threshold voltage Vth (gate-emitter voltage Vge) of IGBTs using SiC also changes each time it is measured. Therefore, the threshold voltage Vth (Vgs, Vge) is not suitable as a physical quantity to measure for accurately estimating the temperature of MOSFETs and IGBTs.

[0006] In contrast, the drain-source voltage Vds of a SiCMOSFET exhibits stable temperature dependence, making it suitable as a physical quantity to measure for accurately estimating the temperature of a SiCMOSFET. However, a drawback of SiCMOSFETs is that the drain-source voltage Vds may exhibit the same value at multiple temperatures. In other words, the drain-source voltage Vds of a SiCMOSFET changes in a way that minimizes with temperature changes, and may exhibit the same value at multiple temperatures. On the other hand, the voltage drop due to the various measurement currents mentioned above changes monotonically with temperature changes, exhibiting different values ​​at all temperatures. Therefore, these voltage drop values ​​can be converted to temperature using their temperature dependence. However, since the drain-source voltage Vds of a SiCMOSFET may exhibit the same value at multiple temperatures, it cannot be converted to temperature using its temperature dependence in the usual way. Furthermore, not only the drain-source voltage Vds of a SiCMOSFET, but also semiconductors that exhibit the same voltage drop due to the measurement current at the voltage drop measurement target at multiple temperatures may be developed in the future.

[0007] The present invention was made to solve the above-mentioned problems, and aims to provide a semiconductor device temperature estimation device that can estimate the temperature of a semiconductor device by utilizing the temperature dependence of the voltage drop, even when the voltage drop due to the measurement current of the part to be measured shows the same value at multiple temperatures. [Means for solving the problem]

[0008] To achieve the above objective, a semiconductor device temperature estimation device according to one aspect of the present disclosure comprises: a heating power supply that supplies a heating current to a semiconductor element to be measured in a first semiconductor device under test; a measuring power supply that supplies a measuring current of a predetermined value to a first target area of ​​the semiconductor element under test, which is a target area for measuring the voltage drop of the semiconductor element under test in the first semiconductor device under test; a voltage sensor that detects the voltage drop of the first target area of ​​the semiconductor element under test in the first semiconductor device under test; and a control device, wherein the control device comprises: a heating step of controlling the heating power supply to heat the semiconductor element under test by supplying the heating current to the semiconductor element under test in the first semiconductor device under test; a heat dissipation step of controlling the heating power supply to stop supplying the heating current after the heating step to dissipate heat from the semiconductor element under test; a measuring current supply step of controlling the measuring power supply to supply the measuring current to the first target area of ​​the semiconductor element under test in the first semiconductor device under test, at least in the heat dissipation step; and in the heat dissipation step, via the voltage sensor, the measuring current A control device that performs the following steps: acquiring voltage drop data of the voltage drop of the first measurement target area; and converting the time series data of the voltage drop of the first measurement target area acquired in the voltage drop acquisition step into time series data of the temperature of the first measurement target area using a temperature / voltage drop curve, which has been acquired in advance and represents the change in the voltage drop of the second measurement target area in response to a change in temperature of the second measurement target area, which is a target area of ​​the second semiconductor device under test for measuring the voltage drop of a semiconductor element under test, wherein the second semiconductor device under test has the same configuration as the first semiconductor device under test, the temperature / voltage drop curve is a curve in which the voltage drop of the second measurement target area changes with at least one of a maximum and minimum value as the temperature of the second measurement target area decreases, and the conversion step is a step of converting the time series data of the voltage drop of the first measurement target area into time series data of the temperature of the first measurement target area by utilizing the fact that all the temperatures of the second measurement target area in the temperature / voltage drop curve each correspond to one voltage drop of the second measurement target area. [Effects of the Invention]

[0009] The present invention provides a semiconductor device temperature estimation device that can estimate the temperature of a semiconductor device by utilizing the temperature dependence of the voltage drop, even when the voltage drop due to the measurement current of the part to be measured shows the same value at multiple temperatures. [Brief explanation of the drawing]

[0010] [Figure 1] Figure 1 is a functional block diagram showing an example of the configuration of a semiconductor device temperature estimation device according to Embodiment 1 of this disclosure. [Figure 2] Figure 2 is a circuit diagram showing the contents of the semiconductor device under test in Figure 1. [Figure 3] Figure 3 is a cross-sectional view showing an example of the schematic structure of the N-channel MOSFET shown in Figure 2. [Figure 4] Figure 4 is a schematic diagram illustrating the mechanism by which the temperature dependence of the MOSFET's on-resistance exhibits a minimum value. [Figure 5A] Figure 5A is a schematic graph showing the threshold voltage dependence of the on-resistance of the MOSFET. [Figure 5B] Figure 5B is a schematic graph showing the drain breakdown voltage dependence of the MOSFET's on-resistance. [Figure 6A] Figure 6A is a plan view showing the structure of the semiconductor device under test in Figure 1. [Figure 6B] Figure 6B is a cross-sectional view showing the structure of the semiconductor device under test in Figure 1. [Figure 7] Figure 7 is a flowchart showing the control contents of the control device shown in Figure 1. [Figure 8] Figure 8 is a graph showing the temperature / voltage drop curve for the voltage drop measurement site in a typical semiconductor device under test. [Figure 9] Figure 9 is a graph showing time-series data of the voltage drop due to the measurement current in a typical semiconductor device under test. [Figure 10]FIG. 10 is a graph showing the temperature / drop voltage curve of the drop voltage measurement target site in the measurement target element of the semiconductor device to be measured, which is used in the conversion process of Embodiment 1. [Figure 11] FIG. 11 is a graph showing the time-series data of the voltage across both ends by the measurement current in the measurement target element of the semiconductor device to be measured, where the temperature is estimated in Embodiment 1. [Figure 12] FIG. 12 is a graph showing the time-series data of the temperature of the measurement target element of the semiconductor device to be measured, where the temperature is estimated by the sequential conversion process of Embodiment 1. [Figure 13] FIG. 13 is a graph showing the time-series data of a plurality of comparison temperatures related to the measurement target element of the semiconductor device to be measured, which is used in the matching conversion process of Embodiment 1. [Figure 14] FIG. 14 is a graph showing the time-series data of a plurality of comparison drop voltages related to the measurement target element of the semiconductor device to be measured, which is used in the matching conversion process of Embodiment 1, together with the time-series data of the voltage across both ends of the measurement target element of the semiconductor device to be measured. [Figure 15] FIG. 15 is a graph showing the temperature / drop voltage curve of the drop voltage measurement target site in the measurement target element of the semiconductor device to be measured, which is used in the conversion process of the example. [Figure 16] FIG. 16 is a graph showing the time-series data of the voltage across both ends by the measurement current in the measurement target element of the semiconductor device to be measured, which is actually measured in the example, and the time-series data of the temperature of the measurement target element of the semiconductor device to be measured, where the temperature is estimated by the sequential conversion process. [Figure 17] FIG. 17 is a graph showing the transient thermal resistance characteristics estimated in the example. [Figure 18] FIG. 18 is a graph showing the time-series data of the voltage across both ends by the measurement current in the measurement target element of the semiconductor device to be measured, which is actually measured in the comparative example, and the time-series data of the temperature of the measurement target element of the semiconductor device to be measured, where the temperature is estimated by the conventional conversion method. [Figure 19]Figure 19 is a graph showing the estimated transient thermal resistance characteristics in the comparative example. [Figure 20] Figure 20 is a graph showing the temperature / voltage drop curve of the voltage drop measurement target area in the semiconductor device under test, used in the conversion process of Embodiment 2. [Modes for carrying out the invention]

[0011] A semiconductor device temperature estimation device according to one aspect of the present disclosure comprises: a heating power supply that supplies a heating current to a semiconductor element to be measured in a first semiconductor device under test; a measuring power supply that supplies a measuring current of a predetermined current value to a first target area of ​​the semiconductor element under test, which is a target area for measuring the voltage drop of the semiconductor element under test in the first semiconductor device under test; a voltage sensor that detects the voltage drop of the first target area of ​​the semiconductor element under test in the first semiconductor device under test; and a control device, wherein the control device includes: a heating step of controlling the heating power supply to supply the heating current to the semiconductor element under test in the first semiconductor device under test to heat the semiconductor element under test; a heat dissipation step of controlling the heating power supply to stop supplying the heating current after the heating step to dissipate heat from the semiconductor element under test; a measuring current supply step of controlling the measuring power supply to supply the measuring current to the first target area of ​​the semiconductor element under test in the first semiconductor device under test, at least in the heat dissipation step; and in the heat dissipation step, detecting the voltage drop of the first measuring current via the voltage sensor. A control device that performs a voltage drop acquisition step of acquiring time-series data of the voltage drop of a target part, and a conversion step of converting the time-series data of the voltage drop of the first target part acquired in the voltage drop acquisition step into time-series data of the temperature of the first target part using a temperature / voltage drop curve that has been acquired in advance and represents the change in the voltage drop of the second target part in response to a change in temperature of the second target part, which is a target part of the second semiconductor device under test that is used to measure the voltage drop of a semiconductor element under test, wherein the second semiconductor device under test has the same configuration as the first semiconductor device under test, the temperature / voltage drop curve is a curve in which the voltage drop of the second target part changes with respect to at least one of a maximum and minimum value as the temperature of the second target part decreases, and the conversion step is a step of converting the time-series data of the voltage drop of the first target part into time-series data of the temperature of the first target part by utilizing the fact that all the temperatures of the second target part in the temperature / voltage drop curve each correspond to one voltage drop of the second target part.

[0012] With this configuration, the temperature / voltage drop curve of the second measurement target portion of the second semiconductor device under test is a curve in which the voltage drop of the second measurement target portion changes with at least one of a maximum and minimum value as the temperature of the second measurement target portion decreases. Therefore, in this temperature / voltage drop curve, one voltage drop of the second measurement target portion may correspond to the temperatures of multiple second measurement target portions. For this reason, the voltage drop of the first measurement target portion cannot be converted to the temperature of the first measurement target portion using the temperature / voltage drop curve in the usual way. However, the temperature of all second measurement target portions each corresponds to one voltage drop of the second measurement target portion. Therefore, by using the temperature / voltage drop curve in a way that takes advantage of this, the time-series data of the voltage drop of the first measurement target portion can be converted to time-series data of the temperature of the first measurement target portion. As a result, even if the voltage drop due to the measurement current of the target portion for which the voltage drop is measured shows the same value at multiple temperatures, it is possible to provide a semiconductor device temperature estimation device that can estimate the temperature of the semiconductor device by utilizing the temperature dependence of the voltage drop.

[0013] The conversion step (i) takes a series of measurement times of the time series data of the voltage drop of the first measurement target part as a series of measurement times of the time series data of the temperature of the first measurement target part, and while tracing the time series data of the voltage drop of the first measurement target part from one end to the other, sequentially identifies the temperature on the temperature / voltage drop curve corresponding to the voltage drop at each measurement time of the time series data of the voltage drop of the first measurement target part as the temperature at each measurement time of the time series data of the temperature of the first measurement target part, and in the process, if the voltage drop of the time series data of the voltage drop of the first measurement target part to be converted first or next corresponds to a plurality of temperatures on the temperature / voltage drop curve, if the one end is the measurement start end of the time series data of the voltage drop of the first measurement target part, the highest temperature among the one or more temperatures that have not yet been identified among the plurality of temperatures is identified as the temperature at each measurement time of the time series data of the temperature of the first measurement target part, and the one end is the first measurement target (ii) A sequential conversion step in which, if it is the end of the measurement of the time series data of the voltage drop of the target part, the lowest temperature among the one or more temperatures that have not yet been identified among the plurality of temperatures is identified as the temperature at a series of measurement times in the time series data of the temperature of the first target part, thereby converting the time series data of the voltage drop of the first target part into time series data of the temperature of the first target part; or (ii) a matching conversion step in which a plurality of time series data of comparative temperatures having different heat dissipation time constants are converted into a plurality of time series data of comparative voltage drop using the temperature / voltage drop curve, pattern matching is performed between the time series data of the voltage drop of the first target part and the plurality of time series data of comparative voltage drop, and the time series data of comparative temperature that corresponds to the time series data of comparative voltage drop that is most similar to the time series data of the voltage drop of the first target part is identified as the time series data of the temperature of the first target part.

[0014] Here, in (i), "when one end is the starting end of the measurement of the time-series data of the voltage drop of the first target part," "the highest temperature among the one or more temperatures that have not yet been identified among the multiple temperatures is the temperature at a series of measurement times in the time-series data of the temperature of the first target part" and "when one end is the ending end of the measurement of the time-series data of the voltage drop of the first target part," "the lowest temperature among the one or more temperatures that have not yet been identified among the multiple temperatures" are uniquely determined. This is because the temperature of the first target part during the measurement period of the voltage drop of the first target part decreases monotonically from the start to the end of the measurement of the voltage drop.

[0015] Furthermore, (i) "If the voltage drop in the time-series data of the voltage drop of the first measurement target area to be converted first or next corresponds to a plurality of temperatures on the temperature / voltage drop curve, if one end is the measurement start end of the time-series data of the voltage drop of the first measurement target area, the highest temperature among the one or more temperatures that have not yet been identified among the plurality of temperatures is identified as the temperature at a series of measurement times in the time-series data of the temperature of the first measurement target area, and if one end is the measurement end end of the time-series data of the voltage drop of the first measurement target area, the lowest temperature among the one or more temperatures that have not yet been identified among the plurality of temperatures is identified "Identifying the temperature at a series of measurement times in the time-series data of the temperature of the first target area to be measured" is synonymous with "First, the temperature corresponding to one end of the time-series data of the voltage drop of the first target area to be measured is identified as the temperature at one end of the time-series data of the temperature of the first target area to be measured, and if the voltage drop of the time-series data of the voltage drop of the first target area to be converted corresponds to multiple temperatures on the temperature / voltage drop curve, then the temperature among those multiple temperatures that is closest to the temperature of the time-series data of the temperature of the first target area to be measured, which was identified immediately before, is identified as the temperature at a series of measurement times in the time-series data of the temperature of the first target area to be measured."

[0016] According to this configuration, in the sequential conversion step (i), if the voltage drop of the time-series data of the voltage drop of the first measurement target to be converted next corresponds to multiple temperatures on the temperature / voltage drop curve of the second measurement target, then all of the temperatures of the second measurement target on the temperature / voltage drop curve correspond to the voltage drop of one second measurement target, so the multiple temperatures on the temperature / voltage drop curve are all different from each other. Therefore, as in the above configuration, if the voltage drop in the time-series data of the voltage drop of the first measurement target area to be converted first or next corresponds to a plurality of temperatures on the temperature / voltage drop curve, if one end is the measurement start end of the time-series data of the voltage drop of the first measurement target area, the highest temperature among the one or more temperatures that have not yet been identified among the plurality of temperatures is identified as the temperature at a series of measurement times in the time-series data of the temperature of the first measurement target area. If one end is the measurement end end of the time-series data of the voltage drop of the first measurement target area, the lowest temperature among the one or more temperatures that have not yet been identified among the plurality of temperatures is identified as the temperature at a series of measurement times in the time-series data of the temperature of the first measurement target area. In this way, the time-series data of the voltage drop of the first measurement target area can be converted to time-series data of the temperature of the first measurement target area.

[0017] Furthermore, in the matching conversion step of (ii), the time-series data of the temperature for comparison is converted into time-series data of multiple voltage drops for comparison using the temperature / voltage drop curve of the second measurement target site. In this case, since the temperature of all the second measurement target sites on the temperature / voltage drop curve corresponds to the voltage drop of one second measurement target site, the time-series data of the temperature for comparison can be converted into time-series data of the voltage drop for comparison using the temperature / voltage drop curve in the usual way.

[0018] As a result, even if the voltage drop due to the measurement current at the target part where the voltage drop is measured shows the same value at multiple temperatures, it is possible to specifically provide a semiconductor device temperature estimation device that can estimate the temperature of the semiconductor device by utilizing the temperature dependence of the voltage drop.

[0019] The second semiconductor device under measurement may be the first semiconductor device under measurement.

[0020] With this configuration, the time-series data of the voltage drop at the first measurement target is converted into time-series data of the temperature at the first measurement target using the temperature / voltage drop curve of the first semiconductor device under measurement. This allows for accurate estimation of the semiconductor device temperature by utilizing the temperature dependence of the voltage drop.

[0021] The semiconductor element to be measured in the first semiconductor device under test is a MOSFET using SiC, the first measurement target area is the channel portion of the SiC MOSFET, and the voltage drop at the first measurement target area is the drain-source voltage; and the semiconductor element to be measured in the second semiconductor device under test is a MOSFET using SiC, the second measurement target area is the channel portion of the SiC MOSFET, and the voltage drop at the second measurement target area is the drain-source voltage.

[0022] With this configuration, the temperature of the channel portion of a SiC-based MOSFET can be estimated using the temperature dependence of the voltage drop.

[0023] Specific embodiments of this disclosure will be described below with reference to the drawings. In the following, the same or corresponding elements are denoted by the same reference numerals throughout all drawings, and redundant explanations are omitted. Furthermore, since the following figures are for illustrative purposes of this disclosure, elements unrelated to this disclosure may be omitted, dimensions may be inaccurate due to exaggeration, simplifications may occur, and the forms of corresponding elements may not match across multiple figures. Also, this disclosure is not limited to the following embodiments.

[0024] (definition) In the following embodiments, the "voltage drop (forward voltage drop)" of the semiconductor element under test and the "voltage across both ends Vm" of the semiconductor device under test are synonymous. Since the term "voltage across both ends Vm" is generally used to estimate the temperature Tj of the semiconductor device under test and measure the transient thermal resistance θth(jc), in the following embodiments, the "voltage drop" of the semiconductor element under test and the "voltage across both ends Vm" of the semiconductor device under test are basically referred to as the "voltage across both ends Vm" of the semiconductor device under test.

[0025] (Embodiment 1) [composition] Figure 1 is a functional block diagram showing an example of the configuration of a semiconductor device temperature estimation device 100 according to Embodiment 1 of this disclosure. Figure 2 is a circuit diagram showing the contents of the semiconductor device 8 under measurement in Figure 1.

[0026] Referring to Figures 1 and 2, the semiconductor device temperature estimation device 100 comprises a heating power supply 1, a measuring power supply 2, a voltage sensor 3, and a control device 5.

[0027] The heating power supply 1 supplies a heating current Ih to the semiconductor device under test (hereinafter sometimes referred to as DUT) 8 to heat the DUT 8. The measurement power supply 2 supplies a measurement current Im of a predetermined value to the voltage drop measurement target area 8b of the DUT 8 where the voltage drop of the semiconductor element under test is measured. The voltage sensor 3 detects the voltage Vm across the DUT 8. The control device 5 controls the operation of the heating power supply 1 and the measurement power supply 2 and obtains time-series data Ctj of the temperature Tj of the semiconductor element under test in the DUT 8.

[0028] The positive terminal of the heating power supply 1 is connected to the positive terminal of the DUT8 by wiring W1, and the negative terminal of the heating power supply 1 is connected to the negative terminal of the DUT8 by wiring W2. A reverse current blocking diode 6 is provided in the middle of wiring W1. The heating power supply 1 supplies a heating current Ih to the DUT8 through wiring W1 and wiring W2.

[0029] The positive terminal of the measurement power supply 2 is connected to the positive terminal of the DUT8 by wiring W3, and the negative terminal of the measurement power supply 2 is connected to the negative terminal of the DUT8 by wiring W4. A reverse current blocking diode 7 is provided in the middle of wiring W3.

[0030] The measuring power supply 2 supplies a small measuring current Im to the DUT8 through wiring W3 and wiring W4. While the negative terminal of wiring W3 and the positive terminal of wiring W4 may be connected to wiring W1 and W2 for the heating current Ih, respectively, this is undesirable because it increases the influence of the wiring impedance of wiring W1 and W2, thereby increasing the oscillation of the measuring current Im when the heating current Ih is stopped.

[0031] The positive input terminal of the voltage sensor 3 is connected to the positive terminal of the DUT8 by wiring W5, and the negative input terminal of the voltage sensor 3 is connected to the negative terminal of the DUT8 by wiring W6. The voltage sensor 3 measures the voltage across the DUT8 (hereinafter sometimes simply referred to as "voltage across") Vm through wiring W5 and W6, and sends the measured voltage across Vmd to the control device 5.

[0032] The control device 5 controls the operation of the heating power supply unit 1a of the heating power supply 1, the measurement power supply 2, and the switch unit 1b of the heating power supply 1, respectively, using control signals Sc1 to Sc3. The control device 5 also creates time-series data Cvm (see Figure 11) of the voltage Vm across the DUT8 based on the voltage Vmd across the voltage sensor 3, and uses this to obtain time-series data Ctj (see Figure 12) of the temperature Tj of the DUT8 (more precisely, the semiconductor device under measurement). In other words, the control device 5 estimates the temperature Tj of the DUT8 in this way. The control device 5 may also use this temperature Tj of the DUT8 to determine the transient thermal resistance characteristics of the DUT8 (see Figure 17).

[0033] The following will explain each of these elements in detail.

[0034] <Semiconductor device under test (DUT) 8> Since DUT8 and the semiconductor device under test for the temperature / voltage drop curve Cvt used in the conversion process described later have the same configuration, the following explanation also applies to the semiconductor device under test for the temperature / voltage drop curve Cvt.

[0035] {overview} Referring to Figure 2, DUT8 includes one or more semiconductor elements whose heat generation state is to be understood, i.e., semiconductor elements to be measured. Examples of such semiconductor elements include diodes, MOSFETs, IGBTs, and bipolar transistors. In this embodiment 1, time-series data Ctj of the temperature Tj during the heat dissipation process of DUT8 is estimated. To understand the heat generation state of these semiconductor elements, DUT8 is heated by a heating current Ih, then the heating current Ih is stopped, and thereafter, during the heat dissipation process of DUT8, a small measurement current Im is supplied to DUT8 while the voltage Vm across DUT8 is measured. This voltage Vm is converted to temperature Tj based on its dependence on temperature Tj. Furthermore, DUT8 is placed on, for example, a water-cooled heat sink for heat dissipation.

[0036] {Semiconductor devices for which this disclosure is particularly effective} This disclosure is particularly effective (and produces a remarkable effect) when the semiconductor device under measurement in DUT8 has a temperature dependence of voltage drop, where the voltage drop at the measurement target area (hereinafter sometimes simply referred to as the "measurement target area") changes with at least one of a maximum and a minimum value as the temperature of the measurement target area decreases. For such a measurement semiconductor device, the temperature / voltage drop curve representing the change in voltage drop at the measurement target area with respect to the change in temperature of the measurement target area of ​​the measurement semiconductor device changes with at least one of a maximum and a minimum value as the temperature of the measurement target area decreases. Currently, at least the applicant knows of SiCMOSFETs as such measurement semiconductor devices, and knows Vds, which is the voltage drop of the channel portion, as the voltage drop of such measurement target area 8b. Therefore, in this embodiment 1, a device having one SiCMOSFET is exemplified as DUT8. However, as will be described later, even MOSFETs other than SiCMOSFETs may exhibit the above-described temperature / voltage drop curve depending on their specifications. Furthermore, in the future, semiconductor devices other than SiCMOSFETs may be developed that have a temperature dependence of voltage drop, where the voltage drop at the measurement target changes with at least one of a maximum and minimum value as the temperature of the measurement target decreases. It goes without saying that this disclosure is particularly useful for DUT8 which includes such a semiconductor device as the measurement target semiconductor device.

[0037] Figure 2 shows an N-channel SiCMOSFET as an example. Reference numeral 8a indicates the body diode, and reference numeral 8b indicates the measurement target area. In this case, the measurement target area 8b is the channel (inversion layer). This N-channel SiCMOSFET is connected to a heating power supply 1 and a measurement power supply 2 such that a positive voltage is applied to the drain D and a negative voltage is applied to the source S. Then, a gate voltage Vg that is positive to the source S and higher than the threshold voltage Vth is applied to the gate G, turning the N-channel SiCMOSFET ON. In this state, first, a heating current Ih and a measurement current Im are supplied. As a result, the heating current Ih flows through the N-type channel (measurement target area 8b) of the N-channel SiCMOSFET, and the DUT 8 is heated. Next, the heating current Ih is stopped, and the drain-source voltage Vds of the N-channel SiCMOSFET due to the measurement current Im is measured as the voltage across both ends Vm. In this embodiment 1, since the DUT8 includes an N-channel SiCMOSFET (hereinafter sometimes referred to as "MOSFET" at the end), which is a single semiconductor element to be measured, both the DUT and the MOSFET are denoted by reference numeral 8.

[0038] On the other hand, if the SiCMOSFET is a P-channel SiCMOSFET (not shown), the P-channel SiCMOSFET is connected to heating power supply 1 and measurement power supply 2 such that a positive voltage is applied to the source and a negative voltage is applied to the drain. Then, a gate voltage Vg that is negative to the source and lower than the threshold voltage Vth is applied to the gate, turning the P-channel SiCMOSFET ON. In this state, first, a heating current Ih and a measurement current Im are supplied. As a result, the heating current Ih flows through the P-type channel of the P-channel SiCMOSFET, heating the DUT8. Next, the heating current Ih is stopped, and the drain-source voltage Vds of the P-channel SiCMOSFET due to the measurement current Im is measured as the voltage across both ends Vm.

[0039] {temperature Tj of semiconductor device} The voltage Vm across DUT8, which is the voltage drop due to the measurement current Im of the semiconductor element under measurement, includes the forward voltage drop of the pn junction, such as the forward voltage of the diode and the forward voltage of the body diode of the MOSFET, Vsd, and the threshold voltage Vth of the MOSFET and IGBT, as well as the voltage drop of the channel (inversion layer), such as the drain-source voltage of the MOSFET, Vds, and the collector-emitter voltage of the IGBT, Vce. Therefore, in this disclosure, for convenience, both the temperature obtained by converting the forward voltage drop of the pn junction to the temperature of the pn junction and the temperature obtained by converting the voltage drop of the channel (inversion layer) to the temperature of the channel (inversion layer) are referred to as the "temperature Tj of the semiconductor device".

[0040] {Temperature dependence of MOSFET on-resistance} Figure 3 is a cross-sectional view showing an example of the schematic structure of the N-channel MOSFET 8 shown in Figure 2. Referring to Figure 3, the N-channel MOSFET 8 has a vertical planar structure, for example, as a high-voltage MOSFET. The N-channel MOSFET 8 comprises a body 80 made of SiC. A pair of n + Type regions 81 and 82 are formed. And a pair of n + A pair of p-type regions 83 and 84 are formed so as to surround the type regions 81 and 82, respectively. Furthermore, a central portion protrudes between the pair of p-type regions 83 and 84, and the portion surrounding this central portion is positioned below the pair of p-type regions 83 and 84. - A drift layer of type n (hereinafter sometimes simply referred to as the "drift layer") 86 is formed. Below the drift layer 86 is the base layer of the main body 80. + A molded region 85 is formed. Furthermore, the central part of the surface of the main body 80 is the central part of the drift layer 86, the inner parts of a pair of p-type regions 83 and 84, and a pair of n + An insulating layer 90 is formed to cover the inner portions of the mold regions 81 and 82. A gate electrode 91 is formed on top of the insulating layer 90. + The source electrode 87 is formed so as to cover the outer portion of the type region 81 and the outer portion of the p-type region 83. +A source electrode 88 is formed so as to cover the outer portions of the p-type region 82 and the outer portion of the p-type region 84. On the other hand, a drain electrode 89 is formed on the back surface (the bottom surface of the n + -type region 85) of the main body 80. A pair of source electrodes 87, 88, a drain electrode 89, and a gate electrode 91 respectively constitute a source S, a drain D, and a gate G. When a gate voltage Vg exceeding the threshold voltage Vth is applied between the source S and the gate G, a pair of n-type channel portions (hereinafter, may be simply referred to as "channel portions") 92, 93 are respectively formed in portions directly below the insulating layer 90 of the pair of p-type regions 83, 84. When a gate voltage Vg exceeding the threshold voltage Vth is applied in a state where a drain-source voltage Vds is applied between the source S and the drain D, the N-channel MOSFET 8 is turned on, and a drain current flows from the drain D toward the source S. This drain current flows from the drain electrode 89, in order, through the n + -type region 85, the drift layer 86, the channel portions 92, 93, and the n + -type regions 81, 82 to the source electrodes 87, 88. In the resistance of the current path of this drain current, the resistance of the channel portions 92, 93 and the resistance of the drift layer 86 are dominant.

[0041] FIG. 4 is a schematic diagram for explaining the mechanism by which the temperature dependence of the on-resistance of the MOSFET exhibits a minimum value. Referring to FIG. 4, as shown in the upper graph, the resistance of the channel portion decreases as the temperature of the MOSFET rises. On the other hand, as shown in the middle graph, the resistance of the drift layer increases as the temperature of the MOSFET rises. Therefore, as shown in the lower graph, the on-resistance of the MOSFET changes such that the resistance of the channel portion and the resistance of the drift layer are combined and exhibit a minimum value as the temperature of the MOSFET rises.

[0042] FIG. 5A is a graph schematically showing the threshold voltage dependence of the on-resistance of the MOSFET. When the resistance of the channel portion becomes low, the threshold voltage Vth also becomes low. Therefore, when the threshold voltage Vth of the MOSFET is lowered, as shown in FIG. 5A, the minimum point of the on-resistance of the MOSFET moves to the low-temperature side.

[0043] Figure 5B is a schematic graph showing the dependence of the on-resistance of a MOSFET on its drain breakdown voltage. The drain breakdown voltage of a MOSFET increases as the resistance of the drift layer increases. The reason for this is as follows: When the MOSFET is off, the more the depletion layer expands in the drift layer, the higher the drain breakdown voltage. On the other hand, the thicker the drift layer and the lower the impurity concentration in the drift layer, the easier it is for the depletion layer to expand. Furthermore, the thicker the drift layer and the lower the impurity concentration in the drift layer, the higher the resistance of the drift layer. Therefore, the drain breakdown voltage of a MOSFET increases as the resistance of the drift layer increases.

[0044] SiCMOSFETs are used as high-voltage MOSFETs due to their excellent performance, while their threshold voltage Vth is kept as low as possible to reduce the resistance of the channel. Furthermore, SiCMOSFETs are used in power semiconductor devices in relatively high temperature ranges. Therefore, it is presumed that SiCMOSFETs exhibit temperature dependence in their operating temperature range, with the on-resistance changing with increasing temperature, exhibiting a minimum value (minimum point).

[0045] Since the voltage Vm across MOSFET8 in DUT8 is the product of the on-resistance of MOSFET8 and the measured current Im, this voltage Vm exhibits temperature dependence, changing with a minimum value (minimum point) as the temperature of MOSFET8 increases.

[0046] {Structure of DUT8} Figure 6A is a plan view showing the structure of DUT8 in Figure 1. Figure 6B is a cross-sectional view showing the structure of DUT8 in Figure 1. Figure 6B shows a cross-section of DUT8 cut along the VIB-VIB line in Figure 6A.

[0047] Referring to Figures 6A and 6B, in the DUT8, for example, a MOSFET chip 61 is placed on an insulating substrate 62 consisting of three layers. Four lead frames 63 are connected to the MOSFET chip 61. The entire MOSFET chip 61, the insulating substrate 62, and the bases of the four lead frames are covered with a sealing resin body 64. The sealing resin body 64 is provided with screw holes 65 for attaching the DUT8 to another component (for example, a heat sink) with screws. In this embodiment 1, the semiconductor device temperature estimation device 100 estimates the temperature Tj of the MOSFET channel portion of the MOSFET chip 61, and uses this temperature Tj to determine the transient thermal resistance θth(jc) of the heat conduction path from the MOSFET channel portion of the MOSFET chip 61 to the outer surface (back surface) of the insulating substrate 62.

[0048] <Heating power supply 1> The heating power supply 1 includes a heating power supply unit 1a and a switch unit 1b. The heating power supply unit 1a is not particularly limited and only needs to be able to output a predetermined large current (for example, 20A to 1000A) as the heating current Ih. The heating current Ih is DC. A constant current source is an example of the heating power supply unit 1a. The constant current source is composed of, for example, a switching power supply, but may also be composed of a linear power supply.

[0049] The switch unit 1b includes, for example, a switching element 11 connected to the positive output terminal of the heating power supply unit 1a, and a diode 12 whose anode is connected to the positive output terminal of the heating power supply unit 1a and whose cathode is connected to the negative output terminal. The switching element 11 is turned on and off by a control signal Sc3 from the control device 5. When the switching element 11 is turned off while the switching element 11 and the heating power supply unit 1a are turned on and a heating current Ih is output from the heating power supply 1, the output current of the heating power supply unit 1a flows through the diode 12, so the output of the heating current Ih can be stopped instantaneously. In this embodiment 1, since the heat dissipation characteristics of the DUT 8 are measured, it is important that the output of the heating current Ih can be stopped instantaneously at a desired timing. Note that the switch unit 1b may be omitted in the heating power supply 1. In this case, the supply of the heating current Ih is started or stopped by the starting or stopping of the heating power supply unit 1a, respectively.

[0050] <Measurement power supply 2> The measuring power supply 2 is not particularly limited and only needs to be able to output a small DC measuring current Im. For example, the measuring power supply 2 is composed of a constant current source. The measuring current Im is, for example, 1A.

[0051] <Voltage Sensor 3> The voltage sensor 3 only needs to be able to measure DC voltage. Examples of voltage sensors 3 include resistive voltage sensors and capacitor voltage sensors.

[0052] <Control device 5> The control device 5 controls the operation of the heating power supply unit 1a, the measurement power supply 2, and the switch unit 1b of the heating power supply 1, respectively, by control signals Sc1 to Sc3. It also creates time-series data Cvm of the voltage Vm across the DUT 8 based on the voltage Vmd across the voltage sensor 3, and uses this to obtain time-series data Ctj of the temperature Tj of the semiconductor element under measurement in the DUT 8. The control device 5 may further obtain the transient thermal resistance θth(jc) of the DUT 8 using the time-series data Ctj of the temperature Tj of the semiconductor element under measurement. The control device 5 is composed of, for example, an arithmetic unit having a processor and memory. A predetermined program for realizing the functions of the control device 5 is stored in the memory of this arithmetic unit, and the control device 5 is realized as a functional block by the processor reading and executing this predetermined program. In this case, the arithmetic unit operates as the control device 5. This arithmetic unit can be composed of, for example, a computer, personal computer, microcontroller, MPU, FPGA (Field Programmable Gate Array), PLC (Programmable Logic Controller), etc.

[0053] Herein, the functions of the elements disclosed herein can be performed using circuits or processing circuits, including general-purpose processors, dedicated processors, integrated circuits, ASICs (Application Specific Integrated Circuits), conventional circuits, and / or combinations thereof, configured or programmed to perform the disclosed functions. A processor is considered a processing circuit or circuit because it includes transistors and other circuits. In this disclosure, “circuit” or “part” is hardware that performs the enumerated functions, or hardware programmed to perform the enumerated functions. The hardware may be hardware disclosed herein, or other known hardware that is programmed or configured to perform the enumerated functions. If the hardware is a processor, which is considered a type of circuit, then “circuit” or “part” is a combination of hardware and software, where the software is used to configure the hardware and / or the processor.

[0054] [Operation] Next, the operation of the semiconductor device temperature estimation device 100 configured as described above will be explained with reference to Figures 1 and 7. This operation is performed by the control device 5 controlling the semiconductor device temperature estimation device 100. Figure 7 is a flowchart showing the control contents of the control device 5.

[0055] <Overview> Referring to Figures 1 and 7, first, the control device 5 performs the heating process and starts the measurement current supply process (step S1). Specifically, the control device 5 outputs an ON signal control signal Sc1 to the heating power supply unit 1a of the heating power supply 1. Then, the heating power supply unit 1a outputs a heating current to the switch unit 1b. This heating current flows through the diode 12 in the switch unit 1b and returns to the heating power supply unit 1a. The control device 5 also outputs an ON signal control signal Sc2 to the measurement power supply 2. As a result, the measurement power supply 2 starts supplying the measurement current Im to the DUT 8.

[0056] Next, the control device 5 outputs an ON signal control signal Sc3 to the switch unit 1b of the heating power supply 1. This turns on the switching element 11 of the switch unit 1b, and the heating current from the heating power supply unit 1a is split and flows through the diode 12 and the switching element 11. The heating current Ih output from the switching element 11 is supplied to the DUT 8 via wiring W1 and W2. This heating current Ih is detected by a heating current sensor (not shown) and is feedback-controlled by the heating power supply unit 1a to a predetermined current value. The control device 5 then starts measuring the voltage drop (forward voltage drop VF) at the measurement target area 8b of the DUT 8. This forward voltage drop VF is used to determine the transient thermal resistance θth(jc).

[0057] Next, the control device 5 starts the heat dissipation process (step S2). That is, once a predetermined heating period has elapsed, the control device 5 stops supplying the heating current Ih to the switch unit 1b of the heating power supply 1, and also stops the heating power supply unit 1a of the heating power supply 1. The "predetermined heating period" is determined appropriately, taking into consideration the structure of the DUT8, the temperature Tj, and the measurement conditions of the transient thermal resistance θth(jc). The "predetermined heating period" is determined to be a sufficient time for the temperature Tj of the DUT8 to saturate. Specifically, once the predetermined heating period has elapsed, the control device 5 outputs an off signal control signal Sc3 to the switch unit 1b of the heating power supply 1. Then, the switching element 11 of the switch unit 1b turns off, and the output of the heating current Ih is stopped. As a result, all the heating current from the heating power supply unit 1a flows to the diode 12. After that, the control device 5 outputs an off signal control signal Sc1 to the heating power supply unit 1a of the heating power supply 1. As a result, the heating power supply unit 1a stops outputting the heating current Ih.

[0058] Next, the control device 5 starts the voltage drop acquisition process (step S3). Specifically, when the supply of the heating current Ih is stopped, a voltage drop occurs across the DUT8 at the measurement target area 8b due only to the measurement current Im. The control device 5 acquires this voltage drop via the voltage sensor 3 and begins measuring it as time-series data Cvm of the voltage Vm across the DUT8.

[0059] Next, the control device 5 terminates the voltage drop acquisition process (step S4) after a predetermined measurement time has elapsed since the heating current Ih was stopped. Specifically, the control device 5 stops acquiring the voltage Vmd across the voltage sensor 3 and creates the final time-series data Cvm of the voltage Vm across the DUT 8.

[0060] Next, the control device 5 terminates the measurement current supply process (step S5). Specifically, the control device 5 outputs an off signal control signal Sc2 to the measurement power supply 2. As a result, the measurement power supply 2 stops supplying the measurement current Im. This effectively terminates the heat dissipation process.

[0061] Next, the control device 5 performs a conversion process (step S6). Specifically, the control device 5 converts the time-series data Cvm of the voltage Vm across the DUT8 into time-series data Ctj of the temperature Tj of the DUT8. The specific procedure for this conversion will be described later. The control device 5 may also further determine the transient thermal resistance θth(jc) using the time-series data Ctj of the temperature Tj of the DUT8 and the forward voltage drop VF.

[0062] Once the conversion process is complete, the control device 5 terminates control of the operation of the semiconductor device temperature estimation device 100.

[0063] <Details of the conversion process> Next, the above-mentioned conversion process will be explained in detail.

[0064] {Target elements of a typical semiconductor device under test} First, we will describe the element to be measured in a typical semiconductor device under test. Figure 8 is a graph showing the temperature / voltage drop curve of the voltage drop measurement site in a typical semiconductor device under test. The element to be measured in a typical semiconductor device under test has a temperature dependence of the voltage across its ends (voltage drop) Vm, as shown in the temperature / voltage drop curve of the voltage drop measurement site in Figure 8. In other words, as is widely known, in a typical semiconductor device under test, the voltage across its ends Vm decreases almost linearly as the temperature Tj of the measurement site increases. Therefore, as in this embodiment 1, in the heat dissipation state, the voltage across its ends Vm increases almost linearly as the temperature Tj of the measurement site decreases in a typical semiconductor device under test.

[0065] Figure 9 is a graph showing the time-series data of the voltage Vm across the element being measured in a typical semiconductor device under test, depending on the measurement current. The temperature Tj of the element being measured in a typical semiconductor device under test decreases monotonically with a certain time constant. Therefore, the voltage Vm across the measurement point of the element being measured in a typical semiconductor device under test increases monotonically with a time constant corresponding to the time constant of the temperature Tj of the element being measured, as shown in Figure 9.

[0066] {Measurement target elements of DUT8 in this disclosure} Next, the measurement target element of the DUT8 in this disclosure will be described. Figure 10 is a semi-logarithmic graph showing the temperature / voltage drop curve Cvt of the voltage drop measurement target area of ​​the semiconductor device under test used in the conversion process of this disclosure. The temperature / voltage drop curve Cvt in Figure 10 is an abstraction of the temperature / voltage drop curve (see Figure 15) measured in the embodiment. In Figure 10, the horizontal axis represents the temperature Tj of the voltage drop measurement target area on a normal scale, and the vertical axis represents the voltage across the voltage drop target area (voltage drop) Vm on a logarithmic scale.

[0067] In this disclosure, DUT8 and the semiconductor device under test for the temperature / voltage drop curve Cvt used in the conversion process may be the same or different. When DUT8 and the semiconductor device under test for the temperature / voltage drop curve Cvt used in the conversion process are different, the former is referred to as the first semiconductor device under test, and the latter as the second semiconductor device under test. Furthermore, the measurement target area of ​​the first semiconductor device under test is referred to as the first measurement target area, and the measurement target area of ​​the second semiconductor device under test is referred to as the second measurement target area. The second semiconductor device under test has the same configuration as the first semiconductor device under test. Regarding the relationship between the first semiconductor device under test and the second semiconductor device under test, for example, the second semiconductor device under test may be a semiconductor device representing a predetermined manufacturing lot, and the first semiconductor device under test may be a semiconductor device belonging to that predetermined manufacturing lot. Based on these facts, if the second semiconductor device under test has the same configuration as the first semiconductor device under test, then even if the second semiconductor device under test differs from the first semiconductor device under test, this does not particularly affect the accuracy of the estimation of the temperature Tj of DUT8.

[0068] For both the first and second semiconductor devices under test, the temperature / voltage drop curve Cvt measured for the measurement target area of ​​each semiconductor element under test is a curve in which the voltage across the measurement target area (voltage drop) Vm changes as the temperature Tj of the measurement target area decreases, with at least one of a maximum and minimum value. Below, we will explain the case where the first semiconductor device under test, DUT8, and the second semiconductor device under test are different. The same applies when the first semiconductor device under test, DUT8, and the second semiconductor device under test are the same, so we will omit that explanation. In this specification, when explaining matters common to the first semiconductor device under test (DUT) and the second semiconductor device under test, for convenience, the first semiconductor device under test (DUT) and the second semiconductor device under test may be collectively referred to as "semiconductor devices under test," and the first measurement target area and the second measurement target area may be collectively referred to as "voltage drop measurement target area."

[0069] For the second semiconductor device under test, the temperature / voltage drop curve Cvt is measured in advance for the second measurement target area of ​​the semiconductor device under test. Specifically, the second semiconductor device under test is placed, for example, in a constant temperature chamber, and while changing the temperature Tj, a measurement current Im is supplied to the second measurement target area of ​​the semiconductor device under test, and the voltage across the second measurement target area (voltage drop) Vm is measured. As a result, the temperature / voltage drop curve Cvt is measured for the second measurement target area of ​​the semiconductor device under test.

[0070] In this embodiment 1, both the first and second semiconductor devices under measurement consist of a single N-channel SiCMOSFET 8 (see Figure 2). The part 8b targeted for voltage drop measurement is the N-channel portion (inversion layer) 92, 93 (see Figure 3). Referring to Figure 10, in the temperature / voltage drop curve Cvt measured for this second semiconductor device under measurement, the voltage across the N-channel SiCMOSFET (voltage drop) Vm changes with a minimum value (minimum point) b as the temperature Tj increases. In other words, in this temperature / voltage drop curve Cvt, the voltage across the ends Vm decreases as the temperature Tj decreases from the high-temperature end, reaches its minimum value at the minimum point b, and increases as the temperature Tj further decreases from the minimum point b. Therefore, for most of this temperature / voltage drop curve Cvt, two temperatures Tj correspond to one voltage across the ends Vm.

[0071] Here, if the second semiconductor device under test has a temperature dependence of voltage drop, such that the voltage drop at the measurement site increases almost linearly as the temperature of the measurement site decreases, as shown in the temperature / voltage drop curve of a typical semiconductor device under test in Figure 8, then in that temperature / voltage drop curve, one temperature Tj corresponds to one voltage Vm across the terminals. Therefore, simply (in the usual (mechanically)) replacing the voltage across the terminals of the first measurement site of the first semiconductor device with the temperature corresponding to that voltage on the temperature / voltage drop curve allows us to convert the time-series data of the voltage across the terminals of the first measurement site into time-series data of the temperature of the first measurement site.

[0072] However, in the above temperature / voltage drop curve Cvt, there may be two temperatures Tj corresponding to one voltage Vm across the terminals. If the conversion is done simply in this way, the voltage Vm across the terminals of the first measurement target area of ​​the first semiconductor device may be replaced with the wrong temperature among the two temperatures Tj corresponding to that voltage Vm on the temperature / voltage drop curve Cvt. As a result, the time-series data Cvm of the voltage Vm across the terminals of the first measurement target area cannot be properly converted to the time-series data Ctj of the temperature Tj of the first measurement target area.

[0073] However, the inventors noticed that in the above temperature / voltage drop curve Cvt, the temperature of each second measurement target corresponds to the voltage Vm across one second measurement target, and that the two temperatures Tj on the temperature / voltage drop curve Cvt are different from each other. The inventors then cleverly utilized this characteristic of the above temperature / voltage drop curve Cvt to create a conversion process that appropriately converts the time-series data Cvm of the voltage Vm across the first measurement target 8b into the time-series data Ctj of the temperature Tj of the first measurement target 8b. This conversion process will be described below.

[0074] <<Sequential Conversion Process>> Figure 11 is a semi-logarithmic graph showing the time-series data Cvm of the voltage across the DUT8 (voltage drop) Vm due to the measurement current Im at the element under measurement in this disclosure, where the temperature is estimated. Figure 12 is a semi-logarithmic graph showing the time-series data Ctj of the temperature Tj at the element under measurement in this disclosure, where the temperature is estimated by the sequential conversion process of this disclosure. Figures 11 and 12 are abstract graphs of the graphs measured in the embodiment, respectively. In Figures 11 and 12, the time-series data Cvm of the voltage across the ends Vm and the time-series data Ctj of the temperature Tj are represented as curves on the graph, respectively. In Figure 11, the horizontal axis represents time t on a logarithmic scale, and the vertical axis represents the voltage across the ends Vm on a normal scale. In Figure 12, the horizontal axis represents time t on a logarithmic scale, and the vertical axis represents temperature Tj on a normal scale. In this embodiment 1, regarding the temperature dependence of the voltage Vm across DUT8, within the assumed operating temperature range (heat dissipation temperature range), the voltage Vm at the end of heat dissipation (end of measurement) is lower than the voltage Vm at the start of heat dissipation (start of measurement). Of course, the voltage Vm across DUT8 may also have a temperature dependence in which the voltage Vm at the end of heat dissipation (end of measurement) is higher than the voltage Vm at the start of heat dissipation (start of measurement) within the assumed operating temperature range (heat dissipation temperature range). The conversion process will be described sequentially below with reference to Figures 10 to 12.

[0075] In the sequential conversion process, first, a series of measurement times t (see Figure 11) of the time-series data Cvm of the voltage Vm across the first measurement target area 8b (see Figure 2) are used to determine the series of measurement times t (see Figure 12) of the time-series data Ctj of the temperature Tj of the first measurement target area 8b. In other words, each time t on the time axis of the time-series data Ctj of the temperature Tj of the first measurement target area 8b shown in Figure 12 is associated with each time t on the time axis of the time-series data Cvm of the voltage Vm across the first measurement target area 8b shown in Figure 11.

[0076] Next, while tracing the time-series data Cvm of the voltage Vm across the first measurement target area 8b from one end to the other, the temperatures Tj on the temperature / voltage drop curve Cvt corresponding to the voltage Vm at each measurement time t of the time-series data Cvm of the voltage Vm across the first measurement target area 8b are sequentially identified as the temperatures Tj at each measurement time t of the time-series data Ctj of the temperature Tj of the first measurement target area 8b. In this process, if the voltage Vm across the time-series data Cvm of the voltage Vm across the first measurement target area 8b to be converted first or next corresponds to multiple temperatures Tj on the temperature / voltage drop curve Cvt, and if the one end is the measurement start end of the time-series data Cvm of the voltage Vm across the first measurement target area 8b, the highest temperature Tj among the one or more temperatures Tj that have not yet been identified is assigned to the time-series data Ctj of the temperature Tj of the first measurement target area 8b at each measurement time t. The temperature Tj is identified, and if one end is the end of the measurement completion side of the time series data Cvm of the voltage Vm across the first measurement target part 8b, the lowest temperature Tj among the one or more unidentified temperatures Tj is identified as the temperature Tj at a series of measurement times t in the time series data Ctj of the temperature Tj of the first measurement target part 8b, thereby converting the time series data Cvm of the voltage Vm across the first measurement target part 8b into time series data Ctj of the temperature Tj of the first measurement target part 8b.

[0077] Specifically, in this case, the aforementioned end is the first end E1, which is the starting end of the time-series data Cvm of the voltage Vm across the first measurement target area 8b. In this case, first, the voltage VmE1 across the first end E1 of the time-series data Cvm of the voltage Vm across the first measurement target area 8b in Figure 11 is converted. In the temperature / voltage drop curve Cvt of the second measurement target area in Figure 10, the voltage VmE1 across the ends corresponds to the temperature Tje1 corresponding to point e1 and the temperature Tjc corresponding to point c. However, the temperature Tje1 corresponding to the first end E1 of the time-series data Cvm of the voltage Vm across the first measurement target area 8b is identified as the temperature Tje1 at measurement time t1 of the time-series data Ctj of the temperature Tj of the first measurement target area 8b. The reason for this is as follows.

[0078] In this case, neither of the two temperatures Tje1 and Tjc has yet been identified as the temperature Tj at a given measurement time t in the time-series data Ctj of the temperature Tj of the first measurement target area 8b. Furthermore, since the first end E1 of the time-series data Cvm of the voltage Vm is the end on the measurement start side of the voltage Vm, the temperature Tj of the first measurement target area 8b at this first end E1 is the highest during the measurement period of the voltage Vm. Therefore, the temperature Tj on the temperature / voltage drop curve Cvt corresponding to the first end E1 of the time-series data Cvm of the voltage Vm is uniquely identified as the higher of the two temperatures Tje1 and Tjc, Tje1. This is the reason for the above. Also, point e1 is identified on the temperature / voltage drop curve Cvt as the point corresponding to the first end E1 of the time-series data Cvm of the voltage Vm. This identifies the temperature Tje1 corresponding to point e1 of the temperature / voltage drop curve Cvt as the temperature Tje1 at time t1 corresponding to the first end E1 of the time series data Cvm of the voltage Vm across both ends of the time series data Ctj of the temperature Tj of the first measurement target area 8b in Figure 12.

[0079] This determines (defines) the first endpoint CE1 of the time-series data Ctj of the temperature of the first measurement target site 8b.

[0080] Next, the time-series data Cvm of the voltage Vm across the first measurement target area 8b in Figure 11 is transformed from the point immediately after the first end E1 to the point immediately before the minimum point B. In this case, the voltage Vm across the first measurement target area 8b in the time-series data Cvm to be transformed next corresponds to multiple temperatures Tj on the temperature / voltage drop curve Cvt. For example, at point A, the voltage VmA across the second measurement target area Cvt in Figure 10 corresponds to temperature Tja corresponding to point a and temperature Tje2 corresponding to point e2. In this case, since neither temperature Tja nor temperature Tje2 has yet been identified as temperature Tj at a series of measurement times t in the time series data Ctj of the temperature of the first measurement target area 8b, the highest temperature Tja among temperature Tja and temperature Tje2 is identified as temperature Tja at time t2, which corresponds to point A of the time series data Cvm of the voltage Vm across the terminals in the time series data Ctj of the temperature Tj of the first measurement target area 8b in Figure 12.

[0081] This determines the CA point of the time-series data Ctj of the temperature Tj of the first measurement target area 8b. Similarly, the interval between the first end CE1 and the CB point of the time-series data Ctj of the temperature Tj of the first measurement target area 8b in Figure 12 is determined.

[0082] Next, the time-series data Cvm of the voltage Vm across the first measurement target area 8b in Figure 11 is transformed at point B. In this case, since point B is a local minimum, the voltage VmB across point B corresponds only to the temperature Tjb corresponding to point b on the temperature / voltage drop curve Cvt. Therefore, this temperature Tjb is identified as the temperature Tjb at time t3 in the time-series data Ctj of the temperature Tj of the first measurement target area 8b in Figure 12, which corresponds to point B in the time-series data Cvm of the voltage Vm across the points.

[0083] This determines the CB point of the time-series data Ctj of the temperature Tj of the first measurement target site 8b.

[0084] Next, the time-series data Cvm of the voltage Vm across the first measurement target area 8b in Figure 11 is transformed from the point immediately following point B to the second end E2. In this case, the voltage Vm across the first measurement target area 8b in the time-series data Cvm to be transformed next corresponds to multiple temperatures Tj on the temperature / voltage drop curve Cvt. For example, at the second end E2, in the temperature / voltage drop curve Cvt of the second measurement target area in Figure 10, the voltage VmE2 across the first measurement target area corresponds to temperature Tje2 corresponding to point e2 and temperature Tja corresponding to point a. In this case, temperature Tja has already been identified as temperature Tja at measurement time t2 of the time series data Ctj of temperature Tj of the first measurement target area 8b, and only temperature Tje2 has not yet been identified as temperature Tj at a series of measurement times t of the time series data Ctj of temperature Tj of the first measurement target area 8b. Therefore, the highest temperature TjE2 of this one temperature Tje2 is identified as temperature Tje2 at time t4, corresponding to the second end E2 of the time series data Cvm of the voltage Vm across both ends of the time series data Ctj of temperature Tj of the first measurement target area 8b in Figure 12.

[0085] This determines the second end E2 of the time-series data Ctj of the temperature Tj of the first measurement target site 8b. Similarly, the interval between CB and the second end CE2 of the time-series data Ctj of the temperature Tj of the first measurement target site 8b in Figure 12 is determined.

[0086] In this way, the time-series data Cvm of the voltage Vm across the first measurement target area 8b is appropriately converted by the sequential conversion process into time-series data Ctj of the temperature Tj of the first measurement target area 8b. In the above conversion process, the time-series data Cvm of the voltage Vm across the first measurement target area 8b was traced from the first end E1 to the second end E2, but conversely, the time-series data Cvm of the voltage Vm across the first measurement target area 8b may be traced from the second end E2 to the first end E1. In this case, if the voltage Vm across the first measurement target area 8b time-series data Cvm to be converted first or next corresponds to multiple temperatures Tj on the temperature / voltage drop curve Cvt, the lowest temperature Tj among the one or more temperatures Tj that have not yet been identified is identified as the temperature Tj at a series of measurement times t in the time-series data Ctj of the temperature Tj of the first measurement target area 8b.

[0087] Furthermore, if the voltage Vm across the terminals of the time-series data Cvm of the voltage Vm across the terminals of the first measurement target part 8b to be converted first or next corresponds to multiple temperatures Tj on the temperature / voltage drop curve Cvt, and if one end is the measurement start end (first end E1) of the time-series data Cvm of the voltage Vm across the terminals of the first measurement target part 8b, then the highest temperature Tj among the one or more temperatures Tj that have not yet been identified is identified as the temperature Tj at a series of measurement times t in the time-series data Ttj of the temperature Tj of the first measurement target part 8b, and the one The process described as follows: "If the end is the end (second end E2) of the time-series data Cvm of the voltage Vm across the first measurement target area 8b, the lowest temperature Tj among the one or more unspecified temperatures Tj among the multiple temperatures Tj is identified as the temperature Tj at a series of measurement times t in the time-series data Ctj of the temperature Tj of the first measurement target area 8b, thereby converting the time-series data Cvm of the voltage Vm across the first measurement target area 8b into the time-series data Ctj of the temperature Tj of the first measurement target area 8b." can be described in various ways. However, any description using these various expressions is essentially the same as the above process if it corresponds to the above process. For example, the above process is equivalent to the following description: "First, identify the temperature Tje1 or Tje2 corresponding to one end (first end E1 or second end E2) of the time-series data Cvm of the voltage Vm across the first measurement target area 8b as the temperature Tje1 or Tje2 at one end (first end CE1 or second end CE2) of the time-series data Ctj of the temperature Tj of the first measurement target area 8b, and then, if the voltage vm across the time-series data Cvm of the voltage Vm across the first measurement target area 8b to be converted corresponds to multiple temperatures Tj on the temperature / voltage drop curve Cvt, identify the temperature Tj that is closest to the temperature Tj of the time-series data Ctj of the temperature Tj of the first measurement target area 8b that was identified immediately before, as the temperature Tj at a series of measurement times t of the time-series data Ctj of the temperature Tj of the first measurement target area 8b."

[0088] <<Matching and Conversion Process>> Figure 13 is a semi-logarithmic graph showing time-series data Ctj1 to Ctj3 of multiple comparison temperatures Tj for the DUT8 under measurement, used in the matching conversion process of this disclosure. Figure 14 is a semi-logarithmic graph showing time-series data M1 to M3 of multiple comparison voltages Vm across the DUT8 under measurement, used in the matching conversion process of this disclosure, together with time-series data Cvm of the voltage Vm across the DUT8 under measurement. Figures 13 and 14 are conceptual diagrams and have been exaggerated and simplified to facilitate understanding of this disclosure. In Figures 13 and 14, the time-series data Ctj1 to Ctj3 of the comparison temperatures Tj and the time-series data M1 to M3 of the comparison voltages Vm across the DUT8 are represented as curves on the graph, respectively. In Figures 13 and 14, the horizontal axis represents time t on a logarithmic scale, and the vertical axis represents temperature Tj on a normal scale, respectively.

[0089] In the matching conversion process, first, multiple time-series data of comparative temperature Tj, Ctj1 to Ctj3, are created for the first measurement target area 8b of the element to be measured in the DUT8. These multiple time-series data of comparative temperature Tj, Ctj1 to Ctj3, can be created to have arbitrary waveforms. In this case, preferred time-series data of comparative temperature Tj, Ctj1 to Ctj3, can be created by utilizing experiments, past data, simulations, etc. Here, for example, these multiple time-series data of comparative temperature Tj, Ctj1 to Ctj3, are created as temperature decrease curves, each having multiple (in this case, three) time constants that are appropriately discretely distributed around a predetermined time constant of heat dissipation in the DUT8 (see Figure 13).

[0090] Next, these multiple time-series data of comparative temperatures Tj, Ctj1 to Ctj3, are converted into multiple time-series data of comparative voltages Vm, M1 to M3, using the temperature / voltage drop curve Cvt of the second measurement target area of ​​the second semiconductor device under test (see Figure 14). In this case, since one voltage Vm corresponds to one temperature Tj in the temperature / voltage drop curve Cvt, the multiple time-series data of comparative temperatures Tj, Ctj1 to Ctj3, can be easily converted into multiple time-series data of comparative voltages Vm, M1 to M3, respectively, by simply (mechanically) replacing the temperature Tj of the first measurement target area 8b of the first semiconductor device under test with the voltage Vm corresponding to that temperature Tj on the temperature / voltage drop curve Cvt. This is an important point in this disclosure.

[0091] Next, pattern matching is performed between the time-series data Cvm of the voltage Vm across the element being measured in DUT8 and the time-series data M1 to M3 of multiple comparison voltage Vm values. At this time, among the multiple time-series data M1 to M3 of the voltage Vm across the elements being measured, the time-series data of the comparison voltage Vm (here, M2) that is most similar to the time-series data Cvm of the voltage Vm across the elements being measured in the first measurement target area 8b, and the time-series data of the comparison temperature Tj (here, Ctj2) corresponding to this time-series data Ctj of the temperature Tj of the first measurement target area 8b is identified as the time-series data Ctj of the temperature Tj of the first measurement target area 8b.

[0092] In this way, the time-series data Cvm of the voltage Vm across the first measurement target area 8b is suitably converted by the matching conversion step into time-series data Ctj of the temperature Tj of the first measurement target area 8b.

[0093] Furthermore, pattern matching between the time-series data Cvm of the voltage Vm across the element being measured in DUT8 and the time-series data M1~M3 of multiple comparison voltage Vm across elements may be performed using text strings instead of curves (images).

[0094] [Examples] To confirm the effects of this embodiment 1, the embodiment was carried out together with a comparative example. The embodiment is the semiconductor device temperature estimation apparatus 100 of this embodiment 1. In the embodiment, the second semiconductor device to be measured was the same as DUT8. In addition, the time-series data Cvm of the voltage Vm across the first measurement target area 8b was converted to the time-series data Ctj of the temperature Tj of the first measurement target area 8b by a sequential conversion process. In the comparative example, the time-series data Cvm of the voltage Vm across the first measurement target area 8b was converted to the time-series data Ctj of the temperature Tj of the first measurement target area 8b by a conventional conversion method. The other configurations are the same as in the embodiment. The contents of the embodiment will be described below with reference to Figures 15 to 19.

[0095] {Temperature / Voltage Drop Curve Cvt} Figure 15 is a semi-logarithmic graph showing the temperature / voltage drop curve Cvt of the voltage drop measurement target area 8b of the semiconductor device under test (DUT8 = second semiconductor device under test) used in the conversion process of the embodiment. In Figure 15, the horizontal axis represents the temperature Tj (°C) of the voltage drop measurement target area 8b on a normal scale, and the vertical axis represents the voltage across both ends Vm (V) on a logarithmic scale. The black circles represent the measurement points. Figure 15 shows the measured data. Referring to Figure 15, the temperature / voltage drop curve Cvt changes such that the voltage across both ends Vm shows a minimum value as the temperature Tj increases. The threshold voltage of the N-channel SiCMOSFET, which is the element under test, was 20V. The measured current Im (drain current ID) was 3A.

[0096] {Time-series data of voltage Vm across both ends (Cvm) and temperature Tj (Ctj)} Figure 16 is a semi-logarithmic graph showing the time-series data Cvm of the voltage across the DUT8 target element due to the measured current Im, as measured in the embodiment, and the time-series data Ctj of the temperature Tj of the DUT8 target element, which was estimated by the sequential conversion process. Figure 18 is a semi-logarithmic graph showing the time-series data Cvm of the voltage across the DUT8 target element due to the measured current Im, as measured in the comparative example, and the time-series data Ctj of the temperature Tj of the DUT8 target element, which was estimated by the conventional conversion method. In Figures 16 and 18, the horizontal axis represents time t (seconds) on a logarithmic scale, the left vertical axis represents temperature Tj (°C) on a normal scale, and the right vertical axis represents the voltage across the DUT8 Vm (V) on a normal scale. In the examples and comparative examples, the temperature dependence of the voltage Vm across DUT8 showed that, within the assumed operating temperature range (heat dissipation temperature range), the voltage Vm at the end of heat dissipation (at the end of measurement) was lower than the voltage Vm at the start of heat dissipation (at the start of measurement).

[0097] Referring to Figure 16, in this embodiment, first, time-series data Cvm of the voltage Vm across the DUT8 was acquired. In this time-series data Cvm of the voltage Vm, the voltage Vm decreased over time, showing a local minimum. Next, this time-series data Cvm of the voltage Vm was converted into time-series data Ctj of the temperature Tj of the DUT8 by the sequential conversion step of Embodiment 1 described above. As a result, appropriately converted time-series data Ctj of the temperature Tj was obtained, as shown in Figure 16.

[0098] On the other hand, referring to Figure 18, in the comparative example, first, time-series data Cvm of the voltage Vm across DUT8 was acquired, similar to the example. In this time-series data Cvm of the voltage Vm across DUT8, similar to the example, the voltage Vm across DUT8 decreased over time, showing a local minimum. Next, this time-series data Cvm of the voltage Vm across DUT8 was converted to time-series data Ctj of the temperature Tj of DUT8 using a conventional conversion method. As a result, as shown in Figure 18, time-series data Ctj of temperature Tj was obtained, in which errors were noticeable near the local minimum.

[0099] {Transient thermal resistance characteristics} Figure 17 is a log-log graph showing the transient thermal resistance characteristics obtained in the example. Figure 19 is a log-log graph showing the transient thermal resistance characteristics obtained in the comparative example. In Figures 17 and 19, the horizontal axis represents time t (seconds) on a logarithmic scale, and the vertical axis represents transient thermal resistance θth (jc) (°C / W) on a logarithmic scale.

[0100] Referring to Figure 17, the embodiment obtained appropriate transient thermal resistance characteristics. On the other hand, referring to Figure 19, the comparative example obtained transient thermal resistance characteristics with noticeable errors in the region corresponding to the vicinity of the minimum value of the time-series data Ctj of temperature Tj in Figure 18.

[0101] {summary} The above embodiments demonstrate that, according to the semiconductor device temperature estimation device 100 of this embodiment 1, even if the temperature / voltage drop curve Cvt of the voltage drop measurement target part 8b is a curve in which the voltage Vm across both ends shows a minimum value as the temperature Tj increases, the temperature Tj of the semiconductor device 8 can be estimated, and consequently, the transient thermal resistance characteristics can be obtained.

[0102] As is clear from the above explanation, according to the semiconductor device temperature estimation device 100 of this embodiment 1, even if the voltage drop (voltage across both ends Vm) due to the measurement current Im of the voltage drop target part 8b shows the same value at multiple temperatures Tj, it is possible to estimate the temperature Tj of the semiconductor device 8 under measurement by utilizing the temperature dependence of the said voltage drop (voltage across both ends Vm).

[0103] (Embodiment 2) Embodiment 2 of this disclosure illustrates an embodiment in which the temperature / voltage drop curve of the voltage drop measurement target portion of the semiconductor device under measurement is a curve that changes such that it shows local minimums and local maximums as the temperature Tj increases.

[0104] Figure 20 is a graph showing the temperature / voltage drop curve Cvt of the voltage drop measurement target portion of the semiconductor device under test, used in the conversion process of Embodiment 2. In Figure 20, the horizontal axis represents temperature Tj, and the vertical axis represents voltage drop (voltage across both ends) Vm. Figure 20 is a conceptual diagram of the temperature / voltage drop curve Cvt of a semiconductor device that may be developed in the future.

[0105] Referring to Figure 20, in Embodiment 2, the temperature / voltage drop curves Cvt of the first and second semiconductor devices under test change to show minimum and maximum values ​​as the temperature Tj increases.

[0106] However, Embodiment 2 differs from Embodiment 1 only in that, in Embodiment 1, a maximum of two temperatures Tj correspond to one voltage Vm across the terminals in the temperature / voltage drop curve Cvt, whereas in Embodiment 2, a maximum of three temperatures Tj correspond to one voltage Vm across the terminals in the temperature / voltage drop curve Cvt. The sequential conversion step and matching conversion step of Embodiment 1 can convert the time-series data Cvm of the terminal voltages Vm to the time-series data Ctj of the temperatures Tj, even if two or more temperatures Tj correspond to one voltage Vm across the terminals in the temperature / voltage drop curve Cvt. Clearly, in Embodiment 2, as in Embodiment 1, the time-series data Cvm of the terminal voltages Vm can be converted to the time-series data Ctj of the temperatures Tj.

[0107] Therefore, according to the semiconductor device temperature estimation device of Embodiment 2, even if the voltage drop (voltage across both ends Vm) due to the measurement current at the voltage drop measurement site shows the same value at multiple temperatures, it is possible to estimate the temperature Tj of the semiconductor device under measurement by utilizing the temperature dependence of the voltage drop.

[0108] (Other embodiments) In Embodiment 2, the first semiconductor device under test and the second semiconductor device under test may have a temperature / voltage drop characteristic in which the voltage across the measured current Im at the voltage drop target site (voltage drop) Vm is the same value at temperatures of 3 or higher.

[0109] Many improvements and other embodiments will be apparent to those skilled in the art from the above description. Therefore, the above description should be interpreted as illustrative only. [Industrial applicability]

[0110] The semiconductor device temperature estimation device of the present invention is useful as a semiconductor device temperature estimation device capable of estimating the temperature of a semiconductor device in which the voltage drop is the same at multiple temperatures. [Explanation of Symbols]

[0111] 1 Heating power supply 1a Heating power supply section 1b Switch section 2 Measurement power supply 3. Voltage Sensor 5 Control device 8. Semiconductor device under test (DUT) Cvt temperature / voltage drop curve Time-series data of voltage across Cvm (voltage drop). Ctj temperature time series data Ih heating current Im measurement current Vm: Voltage across the terminals (voltage drop) Tj temperature

Claims

1. A heating power supply that supplies a heating current to the semiconductor element to be measured in the first semiconductor device under measurement, A measurement power supply that supplies a measurement current of a predetermined value to a first measurement target area, which is the area to be measured for measuring the voltage drop of the semiconductor element of the first semiconductor device to be measured, A voltage sensor for detecting the voltage drop at the first measurement target portion of the semiconductor element to be measured in the first semiconductor device to be measured, A control device is provided, The control device is A heating step of controlling the heating power supply to heat the semiconductor element to be measured in the first semiconductor device to be measured by supplying the heating current to the semiconductor element to be measured, After the heating step, the heating power supply is controlled to stop supplying the heating current, thereby dissipating heat from the semiconductor element to be measured; A measurement current supply step, at least in the heat dissipation step, controls the measurement power supply to supply the measurement current to the first measurement target portion of the first semiconductor element to be measured of the first semiconductor device to be measured, In the heat dissipation step, a voltage drop acquisition step is performed to acquire time-series data of the voltage drop at the first target measurement site due to the measurement current via the voltage sensor, A control device that performs a conversion step of converting the time-series data of the voltage drop of the first measurement target area obtained in the voltage drop acquisition step into time-series data of the temperature of the first measurement target area, using a temperature / voltage drop curve, which has been previously acquired and represents the change in the voltage drop of the second measurement target area in response to a change in temperature of the second measurement target area, which is the target area for measuring the voltage drop of the semiconductor element of the second semiconductor device to be measured. The second semiconductor device under test has the same configuration as the first semiconductor device under test, The temperature / voltage drop curve is a curve in which the voltage drop of the second measurement target changes as the temperature of the second measurement target decreases, with at least one of a maximum and a minimum value occurring. A semiconductor device temperature estimation apparatus, wherein the conversion step is a step of converting time-series data of the voltage drop of the first measurement target into time-series data of the temperature of the first measurement target, by utilizing the fact that the temperature of all the second measurement target locations in the temperature / voltage drop curve each corresponds to the voltage drop of one of the second measurement target locations.

2. The aforementioned conversion process, (i) The series of measurement times of the time series data of the voltage drop of the first measurement target area are taken as the series of measurement times of the time series data of the temperature of the first measurement target area, and while tracing the time series data of the voltage drop of the first measurement target area from one end to the other, the temperatures on the temperature / voltage drop curve corresponding to the voltage drop at each measurement time of the time series data of the voltage drop of the first measurement target area are sequentially identified as the temperatures at each measurement time of the time series data of the temperature of the first measurement target area, and in the process, if the voltage drop of the time series data of the voltage drop of the first measurement target area to be converted first or next corresponds to multiple temperatures on the temperature / voltage drop curve, the one end is A sequential conversion step to convert the time series data of the voltage drop of the first target part into time series data of the temperature of the first target part, wherein if the first end is the starting end of the measurement of the time series data of the voltage drop of the first target part, the highest temperature among the one or more temperatures that have not yet been identified among the multiple temperatures is identified as the temperature at a series of measurement times in the time series data of the temperature of the first target part; and if the first end is the ending end of the measurement of the time series data of the voltage drop of the first target part, the lowest temperature among the one or more temperatures that have not yet been identified among the multiple temperatures is identified as the temperature at a series of measurement times in the time series data of the temperature of the first target part. (ii) A matching conversion step of converting a plurality of time-series data of comparative temperatures having different heat dissipation time constants into a plurality of time-series data of comparative voltage drop using the temperature / voltage drop curve, performing pattern matching between the time-series data of the voltage drop of the first measurement target area and the plurality of time-series data of comparative voltage drop, and identifying the time-series data of comparative temperature that corresponds to the time-series data of comparative voltage drop that is most similar to the time-series data of the voltage drop of the first measurement target area among the plurality of time-series data of comparative voltage drop as the time-series data of temperature of the first measurement target area.

3. The semiconductor device temperature estimation apparatus according to claim 1 or 2, wherein the second semiconductor device to be measured is the first semiconductor device to be measured.

4. The semiconductor element to be measured in the first semiconductor device under measurement is a SiC MOSFET, the first measurement target area is the channel portion of the SiC MOSFET, and the voltage drop in the first measurement target area is the drain-source voltage, and A temperature estimation device for a semiconductor device according to any one of claims 1 to 3, wherein the semiconductor element to be measured in the second semiconductor device to be measured is a MOSFET using SiC, the second measurement target area is the channel portion of the SiC MOSFET, and the voltage drop in the second measurement target area is the drain-source voltage.

Images