Reflected Light Based Junction Temperature Measurement Technique And Microscopic Hyperspectral Analysis Of LEDs

With the rapid progress of semiconductor technique,LED has become the fourth-generation lighting source due to many advantages such as high efficiency,energy saving,long life-time,great stability,environmental friendliness,and so on.It has been widely used in interior lighting,environmental landscape lighting,automobile,indicator,background lighting,and other fields.However,LED still faces many bottlenecks in thermal management.Normally,the conversion ratio of LED from electricity to light energy is only about 30%.Most of the energy will be dissipated in the form of heat,which subsequently causes the increase of junction temperature.And the photoelectric properties of the LED will change correspondingly,such as the luminous efficiency,the uniformity and consistency,and the spectrum.Long-term working at an excessively high junction temperature will lead to reduced device reliability and shortened lifetime.The surface of the LED is usually covered by an optics lens or other encapsulating materials,so detecting tools are difficult to contact the surface of the P-N junction or the chip directly.Thus,non-contact methods should be used injunction temperature test.The electrical parameter methods are most widely applied in junction temperature measurement,but those methods are subject to the change of the working state of LEDs,so not suitable for the LEDs in some specific working conditions.Spectroscopy method,infrared thermography method,and thermoreflectance imaging method are accomplished in optical inspection,but accuracy of these methods remains questionable.Therefore,it is necessary to develop a non-contact high-precision junction temperature test method that can be realized without interrupting or changing the working state of LEDs under test.According to above difficulties,this dissertation proposes a junction temperature detection technology based on relative intensity of reflected light.Furthermore,microscopic hyperspectral analysis is performed to measure the spatial distribution of junction temperature.The main jobs are presented as follows:1.A junction temperature test method based on the relative intensity of reflected light is proposed.The method is based on the linear relationship between the temperature and the reflected light intensity on the surface of the material.An LED is used as the incident light source to illuminate the other LED under test.The spectrum of the reflected light is collected and the relative intensity is obtained.The sensitivity coefficient of the temperature can be calibrated to calculate the junction temperature.The reflected light method utilizes a lock-in amplifier and a photomultiplier tube,greatly improving the sensibility of the method.In addition,the principle of selecting incident light source is proposed to avoid semiconductor excitation,which in turn will affect the measurement accuracy.2.The accuracy of the method is verified by the junction temperature measurement experiments of several colored LEDs under different operating currents.Six red,green,and blue LEDs are used in the experiments,two samples for each color.Under six different operating currents,the junction temperatures of each sample are measured by reflected light method and microthermocouple,respectively.The testing results show that reflected light method has high accuracy in general.In addition,the possible causes of the errors arediscussed.3.Microscopic hyperspectral technology is used to analyze the 2D temperature distribution of LED chip by the reflected light.Microscopic hyperspectral technique is introducedto acquire the spatial and spectral information of the LED chip at the same time.Combining with the reflected light method,the reflection spectrum of each pixel can be obtained,so that the correlation coefficient between the reflected light and the temperature can be acquired.Junction temperature measurement experiments in multiple regions of one LED chip are performed.4.A hyperspectral data processing software is developed based on the Matlab visualization environment.Due to the huge amount of microscopic hyperspectral data,the traditional software cannot fulfill the data processing efficiently.A data processing analysis software named 'HyperSpecDataReader' is developed with the functions of reading,cutting,calculating,and drawing of hyperspectral data.