With the rapid development of LED technology, LEDs have become more and more widely used in various aspects of lighting, and gradually began to occupy the dominant position in the lighting market.
As a new type of illumination source, LED has many advantages over traditional light sources. However, it still has some problems in the actual application process, which causes it to fail to meet its theoretical service life and photometric parameters. Therefore, if a high-reliability, long-life LED light source cannot be realized, even if the light efficiency is good, the high maintenance cost will inevitably limit its application in various fields. The life expectancy of incandescent lamps is generally 1000 hours, and the average life of fluorescent lamps is 10,000 hours. According to the life test of white LEDs by Narendran et al. (LRC), the lifetime of LEDs can reach 100,000 hours.
First, the importance of reliability testing of LED lamps
The commercialization of lighting LEDs has been less than 15 years now, and technology is still evolving and improving. The manufacturers also have a large difference in the level of epitaxial to chip-to-package technology. The quality of different batches of LEDs produced by the same manufacturer is also uneven. Coupled with the difference in the design and use of the application products, the life of the lighting LED is completely less than the theoretical 100,000 hours. Therefore, improving the reliability of LEDs is the top priority of LED research. The study of LED reliability is the premise and basis for improving its reliability and service life. The research on the reliability of LED can not only analyze the failure mechanism of LED from the root, but also propose improvement schemes from design, process and use, and provide a comprehensive evaluation of the reliability of LED lighting products. The failure screening and product quality management provide the basis. Therefore, the research of LED reliability has very important academic value and practical value for LED research and LED industry.
The US Department of Energy (DOE) solid-state lighting 2010 technology route also regards the reliability evaluation test and accelerated life test related evaluation and testing technology of LED lighting products as important research directions in recent years, and is supported by American research institutions. Cree, Philips and other lighting giants have also invested huge human and material resources in research on LED reliability. China's Ministry of Industry and Information Technology, the Ministry of Science and Technology, the National Semiconductor Lighting Engineering R&D and Industry Alliance (CSA), and the Guangdong Provincial Department of Science and Technology have also invested a lot of money in recent years to increase support and research on the reliability of LED lighting products. Many universities, research institutes, and standards alliances have proposed plans to develop standards for LED reliability and accelerated life testing. However, due to the lack of reliable research results, various types of testing and certification are still based on LED lamps when evaluating the life of LED lamps. The normal working state ignites, and then measures the light decay.
Second, the status quo of LED lamps reliability and life testing
According to national standards, the performance test methods of traditional fluorescent lamps and energy-saving lamps are tested for 2000h light decay and 6000h lifetime. At present, CQC's LED lighting products energy-saving certification is to measure the light decay of 3000h, 6000h and 10000h to judge the life of LED lamps; while the US Energy Star certification for LED lighting products, the minimum requirement for light attenuation test is 6000 hours. Although these test methods can detect the actual life of the sample, but the test time is too long, the test cost is too high, and can not meet the market demand. Because of the maturity and stability of traditional luminaire technology, LED technology has developed rapidly. The chip and product shape used in the product have changed greatly every six months. After the 6000-hour light decay test is completed, this product may have been marketed. Eliminated, so we must find faster and more scientific LED luminaire reliability testing methods.
If the LED luminaire is regarded as a system, according to the reliability test, we can divide it into three subsystems, namely the light source module (module) subsystem, the drive power subsystem and the interface subsystem. The author believes that some research institutes have proposed that it is not appropriate to accelerate the life of the overall luminaire and then predict the life. The failure principle and acceleration principle of different systems are very different, and the data obtained is not reliable. Therefore, it should be treated separately according to different subsystems. Compared with the control devices of traditional lamps, the driving power subsystem of LED lamps is compared with the power supply of other electronic products on the market. The technical difficulty is not high, and the impact on the reliability of the light source is not as large as that of the traditional lamp controller. Therefore, rigorous testing according to the current national and international standards for LED control devices can basically guarantee its reliability. The interface subsystem can also ensure its reliability through tests such as vibration and interface insertion and removal. The key need to study the LED light source module subsystem.
Third, the cause of failure of the LED light source module
The LED light source module is generally composed of a substrate, a chip, a packaging material (including a phosphor), a lens, and some modules also include a heat sink and a thermal conductive silicone. The popular packaging methods are DOB and cob. Because of the diversity of LED light source module design and The complexity of the composition, so there are many reasons for failure, generally including the following:
1. Encapsulation material degradation
In the process of daily use of LEDs, long-term work will cause the radiation and temperature of the ultraviolet rays generated by the combination of the blue light of the LED and the inter-band radiation in the GaN system to cause the external packaging material of the LED (such as epoxy resin). The large transparency of the optical transparency of many of the polymers therein causes a decrease in the light extraction efficiency of the LED.
For the problem that the degradation of this packaging material causes a decrease in the light-emitting efficiency of the LED, DL Barton et al. have conducted research and experiments. Experiments show that when the ambient temperature of the LED is 95 °C and the drive current is greater than 40 mA, the pn junction temperature of the LED exceeds 145 ° C. This temperature is the critical state for the package material to change color. If the package material is even carbonized under high current conditions, an opaque substance is formed on the surface of the device or a conductive path is formed, causing the device to fail.
2. Contaminant welding
LED contaminant soldering refers to the defect that LEDs are contaminated by droplets, oil, fibers, dust, etc. during the packaging process, resulting in defects caused by poor or partial contact of the solder joints of the LEDs. The largest LED soldering defect.
According to experiments, when the contaminant covers the entire solder joint, a metal-medium-metal structure, also known as a tunnel junction, is formed at the weld. In the process of device luminescence, due to the existence of the tunnel junction, the peak wavelength of the LED chip will be reduced to 60% of the normal wavelength. Therefore, it is necessary to verify the reliability of the LED package soldering defects.
3. Failure caused by solid crystal primer
In the white LED industry, solid-state adhesives such as aerogel resin, silicone resin, and silver glue are used, and each has its own advantages and disadvantages. Epoxy resin: Insulating adhesive has poor thermal conductivity, but high brightness; Silicone insulation: The thermal conductivity of the adhesive is slightly better than that of the epoxy resin, and the brightness is high. However, due to the proportion of silicon content, the silicone resin and the fluorescent resin remain in the side of the solid wafer. When the epoxy resin is combined, the interlayer phenomenon will occur. After the thermal shock, the peeling will cause the dead lamp; the thermal conductivity of the silver paste is better than the former two, which can prolong the life of the LED chip, but the silver glue is light. The absorption is relatively large, resulting in low brightness. For the two-electrode blue light wafer, the amount of glue is strictly controlled when the crystal is fixed with silver glue, otherwise the short circuit is likely to occur, which directly affects the yield of the product. Therefore, for different types of device products, different solid crystal primers should be properly selected to better reduce the device failure caused by them.
4. Phosphor failure
There are many ways to realize white LEDs. The most common and mature one is to use the blue light generated by the LED chip to excite the yellow phosphor, so the material of the phosphor has a great influence on the attenuation of the white LED. The most popular white light phosphors in the market are YAG aluminum garnet phosphors, silicate phosphors, and nitride phosphors. Compared to blue LED chips, phosphor failure can lead to accelerated LED light decay, which reduces LED life. Experiments show that when the temperature is 80 ° C, the excitation efficiency is reduced by 2%, and then recovered after cooling. This short test shows that the increase in LED temperature will cause the performance of the phosphor to decrease, while the LED is long. Working at high temperatures can cause an irreversible decline in phosphors, and the blue-shifting of LED wavelengths is common.
Therefore, a large part of the light attenuation or even loss of light of a white LED is due to the rapid decay of the performance of the phosphor under heat. Therefore, the quality of the phosphor itself has a significant impact on the normal luminescence lifetime of the LED.
5. Failure caused by heat dissipation problems
LEDs are solid-state semiconductor devices, and LED chips have a small surface area, high current density during operation, and are often required to combine multiple LEDs for illumination. The high density of LEDs leads to high heat dissipation of the chip, and the increase in junction temperature leads to a decrease in light output, which accelerates the chip and shortens the life of the device. Table 1 shows the thermal conductivity of several different materials. It can be seen that in the preparation of power LEDs, the most mature and most used sapphire substrate has a thermal conductivity of only 35 to 46 W/(m×K), which is less than 1/4 of that of Si.
If you want to take into account the adverse effects of color drift in practical applications, thermal design also limits the maximum junction temperature. Due to the continuous increase of the input power of LED chips, the packaging technology of these power LEDs has put forward higher requirements. Today, the heat dissipation problem has become a key factor restricting the development of high-power LEDs.
6. Failure caused by defects in LED GaN-based epitaxial materials
Since there is no substrate material compatible with GaN, there are a large number of defects in GaN films in most LED devices. The mismatch ratio of the lattice constant of GaN material to the current mainstream substrate sapphire is 14%, and the dislocation density of GaN material grown on sapphire substrate is 108/cm3 ~ 1010/cm3.
During the preparation of the LED, the defects of the material will adsorb the carriers, thereby forming a non-radiative recombination center in the active layer, increasing the absorption of light, resulting in a decrease in the luminous efficiency of the LED; when the current is sufficiently large The carrier will only recombine with radiation, but this will cause lattice vibration. The thermal motion of the lattice will accelerate the formation of defects and cause degradation of the LED heterojunction. The metal electrode in the device will migrate along the misalignment under the action of electrical stress and thermal stress, thereby forming a low-pass ohmic barrier, which will cause the optical power of the device to decrease and the leakage current to increase. Therefore, improving the quality of the epitaxial material and reducing the defect density in the material can effectively improve the reliability of the LED device.
7. Failure caused by electrostatic damage
The GaN material has a wide band gap of 3.39 eV and high resistivity. Therefore, the electrostatic charge generated by the GaN-based LED chip during its production and transportation is easily accumulated to generate a high electrostatic voltage. The structure of the sapphire-based GaN-based LED device is very resistant to static electricity and is easily broken by the static electricity generated by it. In the absence of static electricity protection, the static electricity generated by the human body easily breaks down the LED partially, and the LED device will cause permanent failure after being electrostatically broken down.
8.P type GaN ohmic contact aging
Meneghesso et al. analyzed the IV characteristics of LED devices before and after degradation in the failure process of GaN. As shown in Figure 1. The parasitic series resistance of stage I in the figure increases and the forward current decreases at the same voltage. Similar to the Phase II, the reverse leakage current in the IV phase is significantly increased, similar to the Phase III. Meneghesso et al. believe that these changes are due to the ohmic contact of the P-GaN transparent conductive film with the metal wire electrode under the influence of large current and heat, resulting in an increase in series resistance, resulting in a current-intensive effect, thereby causing a decrease in luminous efficiency; In the case of large current injection, the defect will increase, eventually leading to an increase in leakage current. Therefore, the ohmic contact of the metal electrode of P-GaN plays an important role in the optical performance of the LED.
In addition to the above reasons, the other failure causes include soldering holes and cracks of the chip and the substrate, yellowing and cracking of the lens, and open circuit and short circuit of the chip.
Fourth, LED lamp accelerated life test method
In order to quickly find the failure point and weak point of the LED light source module, the reliability life test is usually used to conduct reliability research. The so-called accelerated life test is to improve the stress by using the method of increasing the stress without changing the failure mechanism, so as to obtain the data such as the failure rate and the life in the acceleration in a short time, and then calculate the normal state. The amount of reliability characteristics under stress conditions. Under the condition of increasing stress, the physical and chemical changes in the LED can be accelerated, and the structural design and material defects of the device can be quickly exposed, which provides a basis and reference for the structural design and material optimization of the LED light source module. Current stress conditions commonly used include temperature, humidity, vibration and shock, solar radiation (ultraviolet radiation), electromagnetic radiation, atmospheric pressure, chemical substances (corrosive gases), dust, voltage, current, etc. The effective acceleration stress of the light source module is mainly temperature, humidity, current and vibration. The key to the reliability test method of the LED lamp is how to use the combination of stress, application time and application mode. The accelerated life test can be classified into the following types in accordance with the manner in which stress is applied during the test.
Constant stress accelerated life test
The constant stress accelerated life test is to divide the samples into several groups, each of which is tested under a fixed stress. The stresses experienced by the samples during the test remain unchanged, and the stress levels are not less than three. The test has a longer test time and a relatively larger number of samples. However, compared with the other two accelerated life tests, it is the most mature test method. The test equipment is relatively simple, the test conditions are easy to control, and the test results have small errors, so it is widely used. At present, the US Energy Star test of the life of LED lamps is based on this method. It is necessary to test the 6000-hour light decay data of the LED chip at 55 °C, 85 °C and a manufacturer-specified temperature environment, and then test the LED chip in the LED lamp. The temperature in the middle can be used to calculate the life of the LED lamp. However, the time required for the constant stress acceleration method is still too long to meet the needs of the market.
2. Step stress accelerated life test
The step stress accelerated life test is a stepwise increase in the stress experienced by the sample during the test at a certain time interval until the sample is sufficiently degraded. The test was able to observe component failures in a short period of time and only required a set of test samples. However, the time interval between the two sets of stresses is not easy to determine. If the time interval is too short, the transition effect when changing the stress will affect the aging result of the product. If the time interval is too long, there is no essential difference from the constant stress accelerated life test. Moreover, the step-stress accelerated life test to determine the life-stress relationship of the product, the error is relatively large.
At present, a research hotspot is to use the step temperature stress and constant high humidity stress to accelerate the degradation test of the LED light source subsystem to predict its life. Applying this method must be based on the following five assumptions:
(1) The performance degradation experienced by the test sample is irreversible, that is, the performance degradation process is monotonic.
(2) At each accelerated stress level, the failure mechanism and failure mode of the test sample remain unchanged.
(3) The accelerated degradation data of the test samples at different stress levels have the same distribution form, and the pseudo-failure life of the samples obtained by using the performance degradation data should obey the same distribution type at different stress levels.
(4) The test sample has "no memory characteristics", and its residual life is independent of the manner of accumulation, and only depends on the loaded stress level and the accumulated failure portion.
(5) The performance degradation process of a product can be described by a linear or linear expression.
Generally, three step temperature stress levels and one constant humidity stress level are selected. First, the specific temperature stress step is calculated for a long time, and then the confidence, the number of samples, and the like are determined. When the test time reaches the end time, the degradation with a high degree of fitting is selected. The model is fitted with the degraded data, and the pseudo-failure life of the sample at different temperature stress levels under constant high-humidity stress is calculated. Finally, the reliability characteristics such as the light source reliability distribution function under normal stress level are obtained, and the life is calculated. With this method, the general test time is 2000 hours, which can meet the market demand, but the final reliable research results have not been verified by comparison with the actual ignition lamp life.
3. Pre-stress accelerated life test
The sequential stress accelerated life test is that the stress experienced by the sample during the test increases at the same rate of time until the sample is sufficiently degraded. The advantage of this test is that the acceleration efficiency is the highest and the test time is the shortest. However, the stress increases continuously with time during the test. In order to determine the degree of degradation of the component and the stress-time dependence, it is necessary to repeat the test several times at several different stress-time rate of change, so this determines its Statistical analysis is very complicated and the test setup is relatively expensive and therefore less expensive.
4. High accelerated life test
Now an LED reliability enhancement test is based on step stress acceleration and sequential stress acceleration, combined with high stress life test (HALT) for other stress conditions. According to the standard GWM8287, the first step in the high accelerated life test is to perform a step temperature test.
Finally, according to the actual sample conditions, the sequential stress accelerated cycle test is carried out on the basis of the above four steps, and several stress conditions are continuously increased until the product fails. The main purpose of this method is to find the stress limit of the product, the unpredictable life, and the general test time is less than 100 hours.
These four accelerated life tests have their own advantages and disadvantages. In the actual reliability research test, the test methods should be selected according to the characteristics of different test items.
V. Conclusion
At present, LED luminaire reliability and accelerated life testing methods are under study, and different research institutions have their own research directions and opinions. The author believes that the accelerated test time of about 2000 hours is acceptable in the market, and the test cost is not too high. After the end of the stress accelerated test time, in addition to testing the commonly used LED lamps, such as luminous flux and color coordinates, it is recommended to add LED electrical noise power. Spectrum and other reliability indicators that reflect changes in microphysical mechanisms. Because the change of luminous flux is not sensitive, after the accelerated life test, the actual reliability structure and microscopic physical mechanism of some LED lamps have changed, but the luminous flux and color coordinates have not changed significantly. If the test with other reliability indicators can be improved The reliability of the reliability test results.
Woolen Socks, Feather Socks, Cotton Socks
JNY International Trade Co., Ltd , http://www.jny-socks.com