JP2011247735A - Method of measuring in-plane distribution of heat characteristics of metal/ceramic bonded substrate by laser pulse excitation infrared camera measurement method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To measure the in-plane distribution of heat characteristics of a metal/ceramic bonded substrate without forming a blackening film on a surface.SOLUTION: The in-plane distribution of the heat characteristics of the metal/ceramic bonded substrate is measured by measuring the plane distribution of transient temperature time change generated by radiating the high output laser pulse of the output of 100 W or more and the pulse length of 10-500 milliseconds so as to be exposed to the entire main surface of a subject, with a high speed infrared camera of the sampling rate of 0.01 second or lower.

Description

  The present invention relates to a method for measuring thermal characteristics.

  Metal / ceramic bonding substrates are used as substrates for large currents because of their high thermal conductivity. If there is an abnormal part such as peeling or void at the joint between the metal and ceramic, the heat generated from the semiconductor element cannot be released to the heat sink, causing a failure of the semiconductor element. For this reason, it is important to know the in-plane distribution of thermal characteristics.

  The measurement by ultrasonic flaw detection has been frequently used for the joint state of metal and ceramics. The ultrasonic flaw detection method is a method of observing the state of the interface between a metal and a ceramic by submerging the subject and scanning the ultrasonic flaw detection needle, and can detect abnormal portions such as peeling and voids (Patent Document 1). However, in this document, although the diameter and number of voids are measured, the actual heat radiation characteristics are not measured, and the relationship between the diameter and number of voids and the thermal characteristics is unknown. In addition, since the ultrasonic flaw detection method needs to submerge the subject, it is not efficient as a non-destructive inspection method because the subject needs to be dried after the inspection.

  As a method for measuring thermal characteristics, a laser flash method is generally used. However, in this method, in-plane distribution of thermal characteristics cannot be measured. As a method of measuring the in-plane distribution of thermal characteristics, after forming a blackened film on the surface of the subject, heat is instantaneously applied to the surface with a high-power laser pulse, and a surface where the temperature rises transiently with a thermal imaging device from the back A method for measuring the thermal characteristic distribution in the vertical direction by measuring the internal distribution has been proposed (Patent Document 2). It is known that the effect of the thermal resistance of the blackened film becomes relatively large in a material having a high thermal conductivity, and becomes a major error factor in evaluating thermal characteristics (Non-patent Document 1). For this reason, although a method for improving the uniformity of the blackened film has been proposed (Patent Document 3), if carbon or the like is coated on the surface of the specimen, it is not efficient because it needs to be cleaned after the inspection.

  In addition, although a method for measuring the surface distribution of the thermal characteristics of a minute part by scanning a focused laser beam has been proposed (Patent Document 4), it is necessary to form a metal thin film on the surface, There were problems such as being able to measure only the thermal characteristics.

  There has been proposed a method of inspecting a joint portion of an electronic component by receiving infrared rays generated by irradiating a laser pulse so as to cover the entire subject with an infrared camera (Patent Document 5). Printed wiring by irradiating a laser with an output of 24 W for 66.7 milliseconds and shooting in the range of 33.3 to 66.7 milliseconds (first and second frames at a frame rate of 30 Hz) after the start of irradiation. It is exemplified that a void in a soldered portion on a substrate (glass epoxy substrate) is detected. Resin substrates (thermal conductivity: about 1 to 20 W / m · K) and solder (thermal conductivity: about 20 to 70 W / m · K) have low thermal conductivity and slow heat diffusion, and are exemplified in the measurement. Although it is possible to measure the defect portion by this method, the metal / ceramic bonding substrate (thermal conductivity: 150 to 200 W / m · K) has a higher heat diffusion rate than the sampling rate of the infrared camera. Then, the defective part could not be detected.

JP 2007-81429 A JP-A-4-76446 JP 2007-327851 A JP 2000-121585 A Japanese Patent Laid-Open No. 5-52785

Hong, Baba, Niisato, Measurement of thermal diffusivity of low emissivity materials by hemispherical mirror laser flash method, thermophysical properties vol. 11-2 (1997) p. 136

  The problem to be solved is that there has not been a method for measuring in a short time a surface distribution over a wide range of thermal characteristics of a material having high thermal conductivity such as a metal / ceramic bonding substrate.

  The main feature of the present invention is to measure a wide range of in-plane distribution of the thermal characteristics of a subject by laser pulse excitation infrared camera measurement without performing pretreatment on the subject.

In the measurement method used in the present invention, a high-power laser such as a YAG laser, a diode laser, or a fiber laser whose pulse length can be controlled on the order of milliseconds is used as a heat source. The laser preferably has a uniform spatial energy density distribution such as a beam homogenizer. The laser is applied to the entire area to be examined. An infrared camera with a high sampling rate is used as the detector. An infrared camera can obtain the amount of infrared radiation, so the amount of infrared radiation may be converted to temperature. However, if only the in-plane distribution of thermal characteristics is measured, it may not be converted to temperature.

By optimizing the laser output, pulse length, sampling rate of the infrared camera, and measurement angle between the laser and camera and the subject, the reflectivity of the light is high as in metal / ceramic bonding substrates, and the infrared radiation rate is also high. It is possible to observe the surface distribution of the transient thermal properties of an object with a low surface and high thermal conductivity. Transient temperature change by changing the temperature change with time when the laser output is 100 W or more, the pulse length is 10 to 500 milliseconds and the substrate surface is thermally excited, and the infrared camera sampling rate is 0.01 seconds or less. Can be observed. If the laser output is less than 100 W, the surface of the metal / ceramic bonding substrate is not sufficiently heated, so that abnormal portions such as peeling and voids are not sufficiently detected. If the pulse length is less than 10 milliseconds, heating is insufficient, and if it exceeds 500 milliseconds, the surface state of the irradiated surface may change due to overheating. If the sampling rate of the infrared camera is longer than 0.01 seconds, the resulting heat diffusion phenomenon cannot be photographed. The preferable measurement conditions are a laser output of 100 W or more, more preferably 500 W or more, and a camera sampling rate of 0.01 seconds or less, more preferably 0.005 seconds or less.

  By irradiating a laser pulse so that it covers the entire main surface of the subject and observing the distribution of transient temperature changes on the main surface opposite to the irradiated surface with an infrared camera (transmission arrangement), it passes through the inside of the subject. Information on the incoming thermal properties (thermal diffusivity) can be obtained. In addition, by irradiating one main surface of the subject with a laser pulse and observing the distribution of the transient temperature change on the surface (reflection arrangement), information on the thermal characteristics (thermal permeability) of the subject irradiation surface can be obtained. can get. By combining these, it is possible to obtain the thermal characteristic distribution inside the object and the metal / ceramic bonding interface.

The laser light is disposed at an angle of 10 degrees or more from the normal direction of one main surface of the subject. This is because the surface of the object is metal and the reflectance of light is high, so that the reflected light is prevented from flowing back to the laser output port and becoming unstable. In addition, it is preferable that the infrared camera is also arranged at an angle of 10 degrees or more from the direction perpendicular to the subject main surface. This is because the reflectance of the object surface is high, so that the shadow of the camera itself is not reflected. Also, in order to prevent direct light or specularly reflected light from the incident laser from entering the infrared camera, it is better not to place the infrared camera on the optical path of the direct laser light or reflected light. By arranging the laser, subject, and camera in this way, the surface distribution of thermal characteristics is measured without forming a blackening film on the surface of the subject that easily reflects light, such as a metal / ceramic bonding substrate. It becomes possible.

In general, a method of calculating a thermal diffusivity by analyzing a temporal change in temperature caused by pulse heating by a half-time method or the like is employed (Patent Document 2). Alternatively, as in Non-Patent Document 2, a method of calculating the thermal diffusivity and the heat loss parameter Biot number of each part by applying the change with temperature to the equation is used. Since the target is a metal-ceramic bonding material and is not homogeneous, it is difficult to measure the thermal conductivity at each position by applying these methods. However, the obtained temporal change in temperature is Fourier transformed to convert the temporal change in temperature into a phase and amplitude, and a difference in heat conduction can be expressed as a phase delay or amplitude difference. By illustrating the in-plane distribution of the phase or amplitude obtained by Fourier transform, the in-plane distribution of the in-plane thermal characteristics can be obtained. By setting the time domain to be Fourier transformed as the time domain during laser pulse irradiation, information on the thermal characteristics at the time of temperature rise can be obtained, and by setting the time domain after laser pulse irradiation as the time domain, the thermal characteristics at the time of temperature drop Information about
Tetsuya Baba “Sophistication of Data Analysis in Laser Flash Method Calculation and Evaluation of Thermal Diffusivity by Curve Fitting Method” Proceedings of the 17th Japan Thermophysical Properties Symposium (1996) 379

  The laser pulse excitation infrared measurement method of the present invention excites one main surface of a subject with a high-power laser pulse and observes a temporal change in temperature distribution of the main surface opposite to the irradiation surface with an infrared camera. There is an advantage that can measure the surface distribution of a wide range of thermal characteristics.

FIG. 1 is a reverse image of an ultrasonic flaw detection image of a partially peeled object. Usually, ultrasonic flaw detection is measured by a reflection method. Since the present invention is a transmission method, the images are displayed inverted for comparison. FIG. 2 is a thermal characteristic image obtained by measuring the same subject as in FIG. FIG. 3 is a thermal characteristic image obtained by measuring a specimen without peeling by the method according to the present invention.

  In order to solve the above-described various problems, the present invention is capable of measuring the surface distribution of the thermal characteristics of a high thermal conductive member such as a metal / ceramic bonding substrate using a laser pulse and an infrared camera.

In the metal / ceramic bonding substrate measured by the present invention, the ceramic member is AlN, Si ₃ N ₄ , Al ₂ O ₃ or the like, and the metal plate member is aluminum or copper. The ceramic member and the metal member are joined by heating in a vacuum atmosphere or an inert gas atmosphere while applying pressure through an Ag / Cu-based brazing material or Al alloy containing Ti, Zr, Hf, Nb or the like. . In general, the thickness of the ceramic member is 0.2 to 2 mm, the metal plate member is 0.1 to 1 mm, and the bonding layer is 10 to 100 μm. The maximum size is about 10 cm × 10 cm.

When the Al plate was bonded to the AlN ceramic, a part of the bonding material was removed to prepare a specimen in which a part of the ceramic and the metal plate was peeled off. For this subject, an ultrasonic flaw detection image of the subject was obtained with an ultrasonic flaw detector (Hitachi Engineering & Service Co., Ltd. ultrasonic flaw detector). Those without peeling were prepared without removing the bonding material.

[Example 1]
A diode laser (wavelength: 940 nm, rated output: 2500 W) was used as a laser light source, and an SC-5000 series (display pixel number: 320 × 256, full frame rate: 380 Hz) manufactured by FLIR was used as an infrared camera. . These were connected by a PT / vis system manufactured by e / de / vis and analyzed by pulse excitation thermal image software. The laser output was 500 W, the pulse length was 100 milliseconds, and the sampling rate of the infrared camera was 0.003 seconds. The entire surface of the subject was irradiated with laser light, and the laser was placed at an angle of 20 degrees from the perpendicular to the subject surface. An infrared camera was disposed on the back surface of the laser irradiation surface, tilted by 20 degrees from the perpendicular to the subject surface in the direction opposite to the laser light path. A temperature change curve was obtained for each pixel from 0.1 seconds before laser irradiation to 1 second after the completion of laser irradiation.

  As shown in Equation 1 below, f (t), which is a function of time, is obtained by Fourier transform, and F (ω), which is a function of frequency. Further, since F (ω) is obtained as a complex number including a real part and an imaginary part as in Expression 2, it is mathematically known that the amplitude is calculated by Expression 3 and the phase is calculated by Expression 4.

The symbols used in the mathematical formula are as follows.
t: time f (t): time function ω: frequency j: imaginary unit F (ω): frequency function ReF (ω): real part of frequency function ImF (ω): imaginary part of frequency function | F (ω) | : Amplitude α: Phase

The amplitude | F (ω) | and the phase α can be obtained from the temperature change f (t) using general fast Fourier transform and short-time Fourier transform.

The phase and amplitude of each position were calculated by Fourier transforming a temperature curve for 100 milliseconds immediately after the end of laser irradiation. An image is obtained by matching the phase or amplitude at each position.

The image obtained by the method of the present invention is compared with the image obtained by the ultrasonic flaw detector. These images are amplitude images after Fourier transform. The central part of the partially peeled specimen is a bonded part and the peripheral part is a peeled part, and the ultrasonic flaw detection image and the image obtained by the method of the present invention are in good agreement. The amplitudes of the bonded portion and the peeled portion are different, indicating that the heat transfer is different between the bonded portion and the peeled portion. A uniform image was obtained for the unexfoliated specimen. It can be seen that the heat from the laser irradiation surface is uniformly transferred to the back side. The amplitude of the bonded portion was 0.05 to 0.06, and the amplitude of the peeled portion was 0.1 to 0.2. Since the difference in amplitude between the bonded portion and the peeled portion is 0.04 or more, the peeled portion and the bonded portion are separated. Here, when the difference in amplitude between the peeled portion and the joint portion is 0.02 or more, the images can be visually divided. On the other hand, the phase of the bonded portion was 80 to 90 degrees, and the peeled portion was 100 to 140 degrees. Since the phase difference between the bonded portion and the peeled portion is 10 degrees, the bonded portion and the peeled portion are separated. Here, when there is a phase difference of 10 degrees or more between the bonded portion and the peeled portion, it is visually divided.

[Example 2]
A surface distribution image of thermal characteristics was obtained in the same manner as in Example 1 except that the laser output was 100 W. The amplitude of the bonded portion is 0.01 to 0.02, the amplitude of the peeled portion is 0.04 to 0.09, and the difference in amplitude between the bonded portion and the peeled portion is 0.02 or more. The joint was separated. The phase of the bonded portion was 80 to 90 degrees, and the peeled portion was 100 to 130 degrees. Since the phase difference between the bonded portion and the peeled portion is 10 degrees, the bonded portion and the peeled portion are separated.

[Example 3]
A surface distribution image of thermal characteristics was obtained in the same manner as in Example 1 except that the pulse length was set to 10 milliseconds. The amplitude of the bonded portion is 0.01 to 0.03, the amplitude of the peeled portion is 0.05 to 0.12, and the difference in amplitude between the bonded portion and the peeled portion is 0.02 or more. The joint was separated. The phase of the bonded portion was 80 to 90 degrees, and the peeled portion was 100 to 130 degrees. Since the phase difference between the bonded portion and the peeled portion is 10 degrees, the bonded portion and the peeled portion are separated.

[Example 4]
A surface distribution image of thermal characteristics was obtained in the same manner as in Example 1 except that the pulse length was 500 milliseconds. The amplitude of the joined portion is 0.1 to 0.2, the amplitude of the peeled portion is 0.3 to 0.8, and the difference in amplitude between the joined portion and the peeled portion is 0.1 or more. The joint was separated. The phase of the bonded portion was 80 to 90 degrees, and the peeled portion was 100 to 140 degrees. Since the phase difference between the bonded portion and the peeled portion is 10 degrees, the bonded portion and the peeled portion are separated.

[Example 5]
A surface distribution image of thermal characteristics was obtained in the same manner as in Example 1 except that the sampling time of the infrared camera was set to 0.01 seconds. The amplitude of the bonded portion is 0.02 to 0.03, the amplitude of the peeled portion is 0.05 to 0.12, and the difference between the amplitude of the bonded portion and the peeled portion is 0.02 or more. The part and the joint were separated. The phase of the bonded portion was 80 to 90 degrees, and the peeled portion was 100 to 130 degrees. Since the phase difference between the bonded portion and the peeled portion is 10 degrees, the bonded portion and the peeled portion are separated.

[Example 6]
The laser was disposed at a tilt of 10 degrees from the normal to the subject surface, and an infrared camera was disposed at the back of the laser irradiation surface at a tilt of 10 degrees from the normal to the subject surface. To avoid direct laser light, the infrared camera is tilted 20 degrees from the laser light path. A surface distribution image of thermal characteristics was obtained in the same manner as in Example 1 except for the above arrangement. The amplitude of the bonded portion is 0.07 to 0.09, the amplitude of the peeled portion is 0.13 to 0.25, and the difference between the amplitude of the bonded portion and the peeled portion is 0.02 or more. The part and the joint were separated. The phase of the bonded portion was 80 to 90 degrees, and the peeled portion was 100 to 140 degrees. Since the phase difference between the bonded portion and the peeled portion is 10 degrees, the bonded portion and the peeled portion are separated.

[Example 7]
The laser was disposed at a tilt of 40 degrees from the normal to the subject surface, and an infrared camera was disposed at the back of the laser irradiation surface at a tilt of 40 degrees from the normal to the subject surface. In order to avoid direct laser light, the infrared camera is tilted 80 degrees from the laser light path. A surface distribution image of thermal characteristics was obtained in the same manner as in Example 1 except for the above arrangement. The amplitude of the bonded portion is 0.02 to 0.05, the amplitude of the peeled portion is 0.08 to 0.17, and the difference between the amplitude of the bonded portion and the peeled portion is 0.02 or more. The part and the joint were separated. The phase of the bonded portion was 80 to 90 degrees, and the peeled portion was 100 to 140 degrees. Since the phase difference between the bonded portion and the peeled portion is 10 degrees, the bonded portion and the peeled portion are separated.

[Comparative Example 1]
A surface distribution image of thermal characteristics was obtained in the same manner as in Example 1 except that the laser output was 50 W. Noise increases, the amplitude of the bonded portion is 0.005 to 0.008, the amplitude of the peeled portion is 0.01 to 0.03, and the difference in amplitude between the bonded portion and the peeled portion is 0.02 or less. For this reason, it was not possible to separate the peeled portion and the bonded portion. The phase of the bonded portion was 80 to 90 degrees, and the peeled portion was 90 to 130 degrees. Since the phases of the bonding portion and the peeling portion overlap each other, the bonding portion and the peeling portion cannot be separated.

[Comparative Example 2]
A surface distribution image of thermal characteristics was obtained in the same manner as in Example 1 except that the pulse length was 5 milliseconds. The amplitude of the bonded portion is 0.005 to 0.01, the amplitude of the peeled portion is 0.01 to 0.02, and the range of the amplitude of the peeled portion and the bonded portion overlaps, so the peeled portion and the bonded portion Could not be separated. The phase of the bonded portion was 80 to 90 degrees, and the peeled portion was 90 to 120 degrees. Since the phases of the bonding portion and the peeling portion overlap each other, the bonding portion and the peeling portion cannot be separated.

[Comparative Example 3]
A surface distribution image of thermal characteristics was obtained in the same manner as in Example 1 except that the sampling rate of the infrared camera was set to 0.033 seconds. The amplitude of the bonded portion is 0.05 to 0.06, the amplitude of the peeled portion is 0.06 to 0.2, and the range of the amplitude of the peeled portion and the bonded portion overlaps. The joint could not be separated. The phase of the bonded portion was 80 to 90 degrees, and the peeled portion was 90 to 140 degrees. Since the phases of the bonding portion and the peeling portion overlap each other, the bonding portion and the peeling portion cannot be separated.

[Comparative Example 4]
From the data obtained in Example 1, an image of 0.033 seconds after the start of laser irradiation was obtained without Fourier transform. The strength of the peeled portion and the bonded portion was about 10300, and the peeled portion and the bonded strength overlapped, so the peeled portion could not be separated.

Table 1 summarizes the conditions and measurement results of Examples 1 to 7 and Comparative Examples 1 to 3.

  The present invention can be applied to non-destructive measurement of the surface distribution of the thermal characteristics of a material having a high thermal conductivity such as a metal / ceramic bonding substrate without performing a pretreatment such as blackening film formation on the surface.

DESCRIPTION OF SYMBOLS 1 Junction part in ultrasonic flaw detection image of partially peeled specimen 2 Separation part in ultrasonic flaw detection image of partially peeled specimen 3 Junction part 4 in laser pulse excitation infrared camera image of partially peeled specimen Separation part 5 in the laser pulse excitation infrared camera image of the peeled specimen Joint part in the laser pulse excitation infrared camera image of the specimen not peeled

Claims (1)

The pulse length of a high-power laser with an output of 100 W or more is set to 10 to 500 milliseconds, the laser pulse is extended so as to hit one whole main surface of the subject, and irradiated at an angle of 10 degrees or more from the perpendicular to the surface, sampling A high-speed infrared camera with a rate of 0.01 seconds or less is placed at an angle of 10 degrees or more from the normal of the main surface opposite to the irradiation surface, and the surface distribution of the time change of the temperature after laser pulse irradiation is obtained. A time-dependent change in temperature is converted into phase and amplitude by Fourier transform, and the surface distribution of the thermal characteristics of the metal / ceramic bonding substrate is formed on the surface of the specimen by measuring the surface distribution of the phase and amplitude. How to measure without.