Thomas Zahner

Study on thermal performance of high power LED employing aluminum filled epoxy composite as thermal interface material

The demand for high power light emitting diodes (LEDs) has spawned a dramatic revolution in illumination industry. However, the miniaturization of electronic devices especially LEDs have increased the need for highly sophisticated yet cost effective thermal management solutions in order to sustain this technological advancement. This paper elucidates the thermal behaviour of an LED employing metal filled polymer matrix as thermal interface material (TIM) for an enhanced heat dissipation characteristic. Highly thermal conductive aluminium were incorporated in bisphenol A diglycidylether (DGEBA) epoxy matrix to identify the effect of filler to polymer ratio on the thermal performance of high power LEDs. The curing behaviour of DGEBA was optimized by differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA). The dispersion nature of the Al fillers into polymer matrix was verified with Field Emission Scanning Electron Microscope (FESEM). The thermal performance of synthesized Al filled polymer composite as TIM was tested with 5W green LED employing thermal transient measurement technique. Comparing the filler to polymer ratio, the rise in junction temperature for 60 % Al filled composite was higher by 11.0 °C than 50 % Al filled composite at cured state. In addition, it was also observed from the structure function analysis that the total thermal resistance was 10.96 K/W higher for 60 % Al filled composite compared to 50 % Al filled composite. On the other hand, a significant rise of 9.46 °C in the junction temperature between cured and uncured samples of 50 % Al filled polymer TIM was observed and hence the importance of curing process of metal filled polymer composite for effective heat dissipation has been discussed extensively in this work.

Failure mechanisms of gallium nitride LEDs related with passivation

This paper analyzes the thermally-activated failure mechanisms of GaN LEDs under thermal overstress related with the presence of a PECVD SiN passivation layer. It is shown that the properties of the passivation layer can remarkably affect devices reliability during high-temperature stress: degradation mechanisms identified consist in emission crowding and series resistance increase, attributed to the thermally-activated indiffusion of hydrogen from the passivation to the p-layer, and subsequent p-doping compensation

Inline Rth control: Fast thermal transient evaluation for high power LEDs

As junction-to-case thermal resistance RthJC is a primary performance and reliability parameter for high power Light Emitting Diodes (LED) an accurate specification of this value is of paramount importance. Currently thermal transient characterization methods are reserved to research and quality laboratories. Especially the thermal calibration procedure requires an enormous effort of time. Therefore the RthJC specification of a high volume production is based on a statistical approach. However, high test coverage or even a single unit test is desired. This paper presents a method for inline Rth control for high power LEDs. By skipping the conventional thermal calibration procedure the method compares the measured response of the device under test with a completely thermal characterized reference curve of a reference device. It enables to detect variations in thermal interface materials, e.g. failures in the thermal adhesive attach, with sufficient accuracy within some hundred milliseconds testing time. The achieved measurement results verify the applicability of inline Rth control in a high volume production.

Experimental investigations on the offset correction of transient cooling curves of light emitting diodes based on JESD51-14 and simple semi-empirical approximations

Determination of the thermal resistance of high power light emitting diodes by transient thermal measurements is of rapidly growing interest. Due to electrical disturbances at small delay times, correction algorithms like the offset correction described in JESD51-14 are necessary. A simple model based on a mean temperature is presented which gives insight into the physics of this correction algorithm. It both allows for a rough estimate of time intervals where the correction algorithm is applicable in ideal cases, and it can be used to detect the presence of an interface resistance between the substrate and the package. Measurements and numerical simulations reveal that a significant interface resistance between the epi-layer and the substrate leads to a modification of the thermal transient. Offset correction then leads to an error in the determined thermal resistance in the range of several percent depending on the magnitude of the interface resistance. Additionally some simple semi-empirical approximations for the transient cooling curves are given.

Location resolved transient thermal analysis to investigate crack growth in solder joints

An innovative new test method, location resolved transient thermal analysis (LrTTA), is developed based on transient thermal measurement (TTM). LrTTA uses several distinct diodes on a test chip to detect the thermal performance of interfaces and assemblies. The temperature is measured by the forward voltage time dependent at different locations and inhomogeneities in the interface, e.g. cracks, voids and thickness variations, can be resolved. This investigation is necessary for analysis of the failures in solder joints since the local temperature can be strongly vary due to local bad thermal contact. For first experimental application of the method, a silicon thermal test chip with four different located temperature diodes was employed and soldered on an Aluminium Insulated Metal Substrate (Al-IMS) and exposed to temperature cycles. Transient thermal measurements were performed directly after assembly and after specific temperature shock cycle numbers (−40°C/+125°C). After data processing the increase of the thermal impedance of each diode between the initial “0” cycles and “n” cycles was obtained and correlated with crack distribution in the solder joint by X-ray and scanning acoustic microscopy images. In addition, a finite element (FE) model was set up and used to analyze the solder joint with and without voids and also the crack propagation in the solder joint during temperature shock testing.

A comparison study on thermal characterization of high power leds with different ceramic attach adhesives for automotive lighting applications

In general, the increased electrical current used to drive head lamp LEDs has focused more attention on the thermal paths in the packages. Consequently, the increase in the heat flux can cause thermal runaway and catastrophic failures, limiting the full exploitation of the LED chip technology. The present study aims to develop highly thermal conductive composite materials, based on aluminium nitride (AlN) epoxy composites, as ceramic attach thermal adhesive layer for high power LEDs. Two different polymer matrices were employed to compound the AlN filled epoxy composites; bisphenol-A diglycidyl ether (D1) epoxy resin and a cycloaliphatic compound (Cl). Comparing D1 and C1 with the control glue system, it was found that control glue performed the best. D1 performed the worst with highest thermal resistance in the application. The material properties of the glues and the thermal transient measurement of head lamp LEDs are discussed extensively in this paper.

Evaluation of thermal solder joint quality and thermal performance of PCBs by using standard measurement equipment

By using more powerful LEDs a good thermal management is becoming more and more important. Here usually the thermal interconnects have to be in focus. Therefore a simple and rapid measurement method for the thermal check of the solder joint in a high volume production is necessary. A standard thermal resistance measurement is very time consuming due to the calibration of each device. By using thermal characterized reference devices it is possible to skip the time consuming part. Furthermore, a simple methodology for evaluating the thermal performance of the various metal core PCB (MCPCB) materials and construction is required during development. In this paper we will present how we successfully demonstrated that this new method offers an opportunity to determine the thermal performance of a high power QFN LED on different types of isolated metal substrate (IMS) and with a standard SMU [1] (source measurement units).

Study on thermal performance of high power LED employing aluminium filled epoxy composite as thermal interface material

The demand for high power light emitting diodes (LEDs) has spawned a dramatic revolution in illumination industry. However, the miniaturization of electronic devices especially LEDs have increased the need for highly sophisticated yet cost effective thermal management solutions in order to sustain this technological advancement. This paper elucidates the thermal behaviour of an LED employing metal filled polymer matrix as thermal interface material (TIM) for an enhanced heat dissipation characteristic. Highly thermal conductive aluminium were incorporated in bisphenol A diglycidylether (DGEBA) epoxy matrix to identify the effect of filler to polymer ratio on the thermal performance of high power LEDs. The curing behaviour of DGEBA was optimized by differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA). The dispersion nature of the Al fillers into polymer matrix was verified with Field Emission Scanning Electron Microscope (FESEM). The thermal performance of synthesized Al filled polymer composite as TIM was tested with 5W green LED employing thermal transient measurement technique. Comparing the filler to polymer ratio, the rise in junction temperature for 60 % Al filled composite was higher by 11.0 °C than 50 % Al filled composite at cured state. In addition, it was also observed from the structure function analysis that the total thermal resistance was 10.96 K/W higher for 60 % Al filled composite compared to 50 % Al filled composite. On the other hand, a significant rise of 9.46 °C in the junction temperature between cured and uncured samples of 50 % Al filled polymer TIM was observed and hence the importance of curing process of metal filled polymer composite for effective heat dissipation has been discussed extensively in this work.

Experimental and theoretical considerations on the offset correction of transient cooling curves of light emitting diodes based on JESD51-14

Determination of the thermal resistance of high power light emitting diodes by transient thermal measurements is of rapidly growing interest. Due to electrical disturbances at small delay times, correction algorithms like the offset correction described in JESD51-14 are necessary. A simple model based on a mean temperature is presented which gives insight into the physics of this correction algorithm. It both allows for a rough estimate of time intervals where the correction algorithm is applicable in ideal cases, and it can be used to detect the presence of an interface resistance between the substrate and the package. Measurements and numerical simulations reveal that a significant interface resistance between the epi-layer and the substrate leads to a modification of the thermal transient. Offset correction then leads to an error in the determined thermal resistance in the range of several percent depending on the magnitude of the interface resistance.

Inline Rth control

As junction-to-case thermal resistance RthJC is a primary performance and reliability parameter for high power light emitting diodes (LED) an accurate specification of this value is of paramount importance. Currently thermal transient characterization methods are reserved to research and quality laboratories. Especially the thermal calibration procedure requires an enormous effort of time. Therefore the RthJC specification of a high volume production is based on a statistical approach. However, high test coverage or even a single unit test is desired. This paper presents a method for inline Rth control for high power LEDs. By skipping the conventional thermal calibration procedure the method compares the measured response of the device under test with a completely thermal characterized reference curve of a reference device. It enables to detect variations in thermal interface materials, e.g. failures in the thermal adhesive attach, with sufficient accuracy within some hundred milliseconds testing time. The achieved measurement results verify the applicability of inline Rth control in a high volume production.