application brief AB206 replaces AN11496 Reliability Considerations for SuperFlux LEDs LEDs provide solidstate lighting and are therefore extremely durable. If the recommended soldering and operating conditions are followed, SuperFlux LEDs will survive for the life of the vehicle. SuperFlux LEDs have performed well in a number of durability, reliability, and accelerated life tests. Some of these tests are summarized in Lumileds’ SuperFlux LED Reliability Data Sheet. A current revision can be obtained by contacting your local Field Sales Engineer or at the following URL: http://www.Lumileds.com This section is dedicated to communicating: 1) the major tests included in the Reliability Data Sheet, 2) the appropriate testing of assemblies containing SuperFlux LEDs, and 3) providing information regarding the typical change in SuperFlux LED lightoutput over typical application lifetimes. Please note that the Reliability Data Sheet also includes the cumulative sample size and failure rate (if other than zero). Table of Contents Durability, Reliability, & Accelerated Life Tests Mechanical Shock Vibration Operating Life Weatherability Corrosion Resistance Temperature Cycle Simultaneous Power and Temperature Cycle Solder Heat Resistance Strife Test Recommended LED Assembly Validation Tests Verifying Root Causes of Power & Temperature Cycle Failures Changes in Light-Output During the Operating Life of AlInGap SuperFlux LEDs 1 2 2 2 3 3 3 3 4 5 6 6 8 Table 6.1 RELIABILITY, DURABILITY AND ACCELERATED LIFE TESTS Mechanical Shock Vibration Variable Frequency Vibration Operating Life Tests: Condition A: Ta = 100°C Condition B: Ta = 85°C Condition C: Ta = 55°C Condition D: Ta = 25°C Condition E: Ta = 40°C Weatherability Condition A: Forward Condition B: Reverse Salt Atmosphere Corrosion Temperature Cycle Condition A: 55 to 100°C Condition B: 40 to 120°C Condition C: 40 to 100°C Power & Temperature Cycle Condition A: 70 mA Condition B: 50 mA Solder Heat Resistance Strife MILSTD883 Method 2002 Condition B: 3 shocks each XYZ axis, 3000g, 0.3 ms MILSTD883 Method 2007 Condition A: 4 cycles, 4 minutes, each XYZ axis, 20g minimum, 20 to 2000 Hz SAE 575 DEC88 Section 4.1: 10 to 55 Hz, 1 mm peaktopeak, 2 minutes per cycle, 60 minutes total duration All tests of duration 1,000 hours If = 20 mA (AS)/30 mA (TS); Tj ~ 110°C If = 45 mA (AS)/50 mA (TS); Tj ~ 110°C If = 70 mA; Tj ~ 90°C If = 70 mA; Tj ~ 60°C If = 70 mA; Tj ~ 5°C 85°C, 85% RH, 45 mA, 1000 h 85°C, 85% RH, 5 V, 1000 h MILSTD883 Method 1009: 35°C, 48 h SAE 575 DEC88: 25°C, 2 times 24 h with onehour dry MILSTD883 Method 1010: 15minute dwell, 5minute transfer, 300 cycles 15minute dwell, 5minute transfer, 300 cycles 15minute dwell, < 10 s transfer, 300 cycles 40°C to 85°C, 3°C per minute transfer rate, 2hour cycle 70 mA, 5minute on/off, ~ 225°C/W thermal resistance (Tj ~ 120°C), 50 cycles 50 mA, 5minute on/off, ~ 225°C/W thermal resistance (Tj ~ 110°C), 50 cycles Reference Figure 6.3 Durability, Reliability, & Accelerated Life Tests Mechanical Shock Lumileds SuperFlux LEDs are encapsulated in a Condition B to validate the SuperFlux LEDs solid plastic package. Therefore, they are mechanical shock performance. This test was extremely rugged and easily survive even conducted three times in each direction at 3000g extreme mechanical shock tests. As shown in for 0.3 milliseconds. All of the SuperFlux LEDs Table 6.1, Lumileds used the test conditions subjected to this test survived without damage. specified in MILSTD883 Method 2002 Vibration SuperFlux LEDs perform extremely well in Ultrasonic welding is frequently used to secure typical vibration tests (up to 2 kHz). SuperFlux the lamp housing to the lens and is occasionally LEDs have survived MILSTD883 Method 2007 used to secure the printed circuit assembly Condition A and SAE 577 vibration tests. containing the LEDs to mounting posts in the lamp housing. Although ultrasonic welding of 2 the lamp housing to the lens has not produced failures of SuperFlux LEDs have been attributed any known LED failures it has been identified as to ultrasonic welding, the potential for this failure the root cause of some catastrophic failures of mode still exists. For this reason, heatstaking or resistors and other leaded components. In a lowerfrequency joining process such as addition, ultrasonic welding of mounting posts vibration welding is preferred for securing the has been identified as the root cause of some mounting posts of the lamp housing to the PCB. catastrophic failures of LEDs. In these cases, In cases where ultrasonic welding must be used, the transfer of ultrasonic energy from the it is extremely important to build many evaluation mounting posts resulted in a broken wire within units to prove the design’s ability to withstand the immediately adjacent LEDs. Although no ultrasonic vibrations. Operating Life Because SuperFlux LEDs are solidstate validate the operating life performance of the devices, they survive thousands of hours of SuperFlux LEDs. These LEDs perform well over operation. As summarized in Table 6.1, their operating temperature range of –40 °C to Lumileds used several ambient temperatures to 100 °C. Weatherability SuperFlux LEDs also perform well in performance after exposure to humidity. These weatherability tests. As summarized in Table tests were performed with the LEDs mounted on 6.1, Lumileds used both forwardbiased and a PCB. No housing was used for this test. reversebiased humidity tests to validate Corrosion Resistance The metal leadframe used in SuperFlux LEDs is The two tests listed in Table 6.1 were used to made from a copper alloy. The portion of the validate SuperFlux LED performance during leadframe that is inside the plastic body is first corrosion testing. The MILSTD883 test was plated with nickel and then silver. The portion of performed using individual LEDs, while the SAE the leadframe that is outside the plastic body is 576 test was performed using LEDs mounted on coated in tinlead solder. Therefore, SuperFlux a printed circuit board. No housing was used for LEDs perform well in corrosion resistance tests. either test. Temperature Cycle The semiconductor that emits the light in leadframe, and the plastic that forms the LED’s SuperFlux LEDs, the leadframe that it is body all have different coefficients of thermal mounted on, the bond wire that connects the expansion. Therefore, thermal cycle test anode of the chip to the anode of the performance is an important measure of 3 SuperFlux LED durability. One of the two typical avoid these failure mechanisms, it is important wearout mechanisms in thermal cycle testing is that the maximum junction temperature listed in separation between the plastic encapsulant and the SuperFlux LED Technical Data Sheet is not the sides of the reflector cup in the leadframe as exceeded (refer to Application Brief AB204 shown in Figure 6.1. This type of separation lifts Thermal Management Considerations for the semiconductor die off of the leadframe and SuperFlux LEDs). Lumileds used the three causes an open circuit or intermittentopen temperature cycle tests listed in Table 6.1 to circuit. The second typical failure mechanism is validate SuperFlux LED performance after a neckeddown wire break above the ballbond thermal cycle testing. (wire bond) as shown in Figure 6.2. In order to Figure 6.1 Example of LED die delamination. Figure 6.2 Example of neck-down wire break. Simultaneous Power & Temperature Cycle Simultaneously cycling both power and Although SAE J1889 describes a power temperature can accelerate typical LED failure temperature cycle test, Lumileds used more modes. Although these conditions are much extreme power temperature cycle conditions to more extreme than those seen in actual field validate SuperFlux LED performance. SAE J1889 use, this type of test is very helpful in comparing recommends powering the device while the performance of different designs. warmingandhot and leaving it unlit when Performance under these conditions is coolingandcold. Lumileds’ tests, which are benchmarked during initial LED product summarized in Table 6.1, are similar to SAE validation and then checked during subsequent J1889, but the power is cycled by excessive validation tests for process or material changes. solder heat, or improperly manufactured This test should also be used when validating SuperFlux LEDs. Lumileds also extends the test assemblies containing SuperFlux LEDs to from the 25 cycles recommended in SAE J1889 ensure that excessive selfheating will not result to a minimum of 50 cycles. in premature failure of SuperFlux LEDs. 4 Solder Heat Resistance Strife Test SuperFlux LEDs have been designed so that thermal stressing of the LEDs. Lumileds’ heat generated within the diode is quickly recommended soldering profiles are located in dissipated out through the copper leads. This Application Note 11492 Mechanical Design thermally efficient design allows these LEDs to Considerations for SuperFlux LEDs. Lumileds be driven at higher power levels relative to other uses the solder heatresistance strife test shown LEDs, thus resulting in optimal optical in Figure 6.3 to verify that SuperFlux LEDs can performance. However, when the leads are be soldered without damage. The temperature immersed in solder, this design also allows heat cycles and hot functional test ensure detection of to conduct from the leads up into the active latent damage or intermittent connections. This area where the diode is located. Because test sequence was used to validate SuperFlux SuperFlux LEDs transfer heat from soldertodie LED performance. In addition, this test is much more quickly than other LEDs, it is repeated regularly to monitor the performance extremely important that recommended solder of SuperFlux LEDs. profiles are followed during assembly to avoid Table 6.2 Recommended Test Test Parameters PostSoldering 30X Visual Inspection Refer to Application Note 11492 Mechanical Design Considerations for SuperFlux LEDs for a description of the procedure and accept/reject criteria. Ta = 25°C, 500 hours at nominal drive conditions 40°C to 85°C [1], dwell & transfer times selected based on thermal mass of the assembly, 50 cycles 40°C to 85°C [1], 3°C per minute transfer rate, 2hour cycle, 5minute on/off, 50 cycles at nominal drive conditions. Or, Room Temperature Operating Life Temperature Cycle Power Temperature Cycle SAE J1889 OCT93: 40°C to 85°C [1], 0.6°C to 5°C per minute transfer rate, onehour minimum dwell, powered at nominal drive conditions while warming and hot, no power while cooling and cold. Note: These conditions EXCEED Lumileds recommended soldering profiles. Figure 6.3 Solder Heat Resistance Strife Test. Table 6.2 Recommended Validation Testing. 5 Recommended LED Assembly Validation Tests Lumileds recommends that customers use the ambient temperature can reach temperatures in tests listed in the Table 6.2 to validate the excess of 100°C. Lumileds recommends that durability of assemblies containing SuperFlux Lighting System Suppliers work with vehicle LEDs. Other tests (such as vibration, or manufacturers to perform temperature corrosion resistance) may be included to check measurements on similar vehicles or mockups the performance of other components, to determine the actual operating and storage materials, or interconnections in the assembly. conditions to which the LED assembly will be Please realize that Lumileds cannot guarantee exposed. During these studies it is important for LED performance during durability tests if the the Lighting System Suppliers to record both maximum operating or storage temperature temperature and supply voltage measurements. ranges, or the maximum junction temperature Both of these parameters will be required when listed in the SuperFlux LED Technical Data designing assemblies for ‘worst case’ conditions Sheet are exceeded. In some applications, such (reference Application Brief AB203 Electrical as highmount stop lamps mounted in the Design Considerations for SuperFlux LEDs.) headliner behind rear window glass, the local Verifying Root Causes of Power & Temperature Cycle Failures The most common causes for SuperFlux LED as surface mount glue cure, the soldering failures during Power & Temperature Cycle operation, any subsequent rework or touchup testing are excessive heat during soldering, operations, and any postsolder heating or thermal shock, and exceeding the devices cooling cycles such as conformal coating cure. absolute maximum junction temperature. (Reference Application Note 11492 Mechanical Design Considerations for SuperFlux LEDs for Separation between the leadframe and the precautions required when attaching thermo encapsulant is usually caused by excessive couples to the leads of LEDs.) The heating and solder heat or thermal shock. If this type of cooling rates must not exceed ± 3°C per second, failure is suspected, the root cause can be and the preheat and soldering temperatures confirmed by measuring the wave solder station should follow the recommendations provided in temperatures and the temperature gradients Application Note 11492 Mechanical Design that the SuperFlux LEDs were exposed to Considerations for SuperFlux LEDs. during assembly. This is accomplished by attaching a thermocouple to the lead of an LED Exceeding the maximum junction temperature on the solderside of the PCB and recording its of the SuperFlux LEDs during Power & temperature profile during each operation. This Temperature Cycle testing usually results in a should include any preliminary operations such neckeddown wire break above the ballbond on 6 the LED chip (see Figure 6.2). To verify this root After all thermocouples have been attached on cause, the junction temperature during a typical the LED assembly, the LED assembly should be Power & Temperature Cycle should be mounted inside the outer housing. Provisions measured. must be made for the thermocouple leads to exit the assembly for attachment to a monitoring The maximum junction temperature can be device. Place the device in the chamber in the determined by adding the temperature way it would be placed during the Power & difference between the LED lead and the LED Temperature Cycle test. junction to the maximum lead temperature. The temperature difference between the LED lead Heat the chamber to the maximum ambient and the LED junction is the product of the temperature to be used during the test, and forward current flowing through the LED, the power the LED assembly (or assemblies). forward voltage of the LED (measured at this Monitor and record the pin temperatures of the forward current), and the thermal resistance of SuperFlux LED leads until the temperatures the SuperFlux LED (listed in the Technical Data reach steadystate. Sheet). This is expressed as the following equation: Tj = Tp + R θjpVf If The junction temperature can be calculated (6.1) using Equation 6.1. An estimate of the maximum Where: junction temperature can also be obtained by Tj = Junction temperature performing this test on a test bench in normal Tp = Temperature of the LED pin (lead) ambient temperatures. Note that this test may R θjp = Junctiontopin (lead) thermal resistance not be able to duplicate the thermal conditions Vf = Forward voltage (at If) inside the test chamber such as airflow around If = Forward current the printed circuit assemblies and heating from adjacent boards. The ambient temperature of the If the junction temperature from this evaluation room during the test should be noted. Using this exceeds the maximum junction temperature approach, the maximum junction temperature is listed in the SuperFlux LED Technical Data the sum of the measured pin temperature, the Sheet, then it is unlikely that the assembly will temperature rise between the pin and the consistently pass the Power & Temperature junction calculated above, and the difference Cycle test. between the room ambient and the maximum ambient temperature to be used in the Power & The LED pin temperature can be measured by Temperature Cycle chamber. This is expressed attaching a thermocouple to one of the cathode as the following formula: Tj = Tp + (R θjpVf If ) leads of the LED on the underside of the PCB (6.2) + (Tamax – Tatest) on the solder pad; making sure that it has a good thermal contact to the metal lead. LEDs Where: near the center of the assembly and near any Tamax= Maximum ambient temperature from Power & Temperature Cycle test heat sources (such as resistors or other driving Tatest = Ambient temperature of bench test circuitry) should be selected for monitoring. 7 Changes in LightOutput During the Operating Life of AlInGap SuperFlux LEDs The lightoutput of SuperFlux LEDs will variation within a batch and between batches of gradually change during their lifetime. This SuperFlux LEDs. Figure 6.5 shows that the change is usually a gradual reduction in light amount of lightoutput degradation is output; however it is also possible to have a proportional to forward current. This figure also slight increase in lightoutput during early shows that some batches of SuperFlux LEDs get operation followed by a gradual reduction. This slightly brighter before starting the typical change in lightoutput is called lightoutput decline. Ambient temperature, humidity, and degradation. Typical amounts of lightoutput sunlight have minimal effect on lightoutput degradation for some popular automotive degradation as shown in Table 6.4, Table 6.5, applications are summarized in Table 6.3. Light and Figure 6.6. For SuperFlux LEDs, Lumileds is output degradation occurs most quickly during careful to select AlInGaP material with an the initial hours of operation and then slows with average lightoutput degradation of less than time as shown in Figure 6.4. There is some 35% after 1000 hours of operation at 70 mA. Table 6.3 Lifetime Change in Light-Output for SuperFlux LEDs in Automotive Signaling Applications Lifetime LightOutput Change @ Ta = 25°C; If = 20 mA (Tail) and 60 mA (Stop and Turn) Function Operating Life (h) Typical LOP Change Tail Stop Turn 1,500 to 3,000 750 to 2000 200 to 700 5 to 5% 30 to 5% 20 to 5% Figure 6.4 Operating life test results for HPWA-MHOO LEDs driven at 70mA, 55°C. Figure 6.5 Change in light-output over time for HPWT-DHOO driven at multiple currents at 55°C. 8 Light output degradation for SuperFlux LEDs is temperature on the rate of lightoutput a function of the current density in the LED degradation. semiconductor chip. The rate of change will be lower when SuperFlux LEDs are driven at lower Figure 6.6 shows the lightoutput degradation of currents. Figure 6.5 shows how the lightoutput some samples mounted in the Arizona Desert. varies over time for LEDs driven at several These samples were directly exposed to the currents. Although this particular lot of LEDs harsh local weather conditions for over a year. became slightly brighter with time at the lower Some of these samples were mounted on a forward currents, their lightoutput peaked and verticallyoriented, southfacing printed circuit then began the typical logarithmic decline. At board. Other samples were mounted on a high currents SuperFlux LED lightoutput will printed circuit board in a special apparatus that peak within a few hours. At lower currents they collects and concentrates sunlight by a factor of may continue to get brighter for several hundred six. A picture of this sun concentrator is shown in hours before beginning the normal gradual Figure 6.7. reduction over time. As shown in Figure 6.6, the concentrated Ambient temperature has negligible effect on sunlight had only a small effect on device the lightoutput degradation of SuperFlux LEDs. performance. During the initial product qualification, HP monitored the change in lightoutput during the The rate of change in lightoutput will vary within operating and forwardbiased weatherability a batch and from one batch of LEDs to another. tests listed in Table 6.1. The average changes Within a lot, the individual standard deviation for for one of the TS AlInGaP wafers used for these results is approximately 5% after 1000 hours of tests are shown in Tables 6.4 and 6.5. The operation. Lumileds Quality System is designed different stress temperatures and the presence to prevent the use of AlInGaP semiconductor of humidity had negligible effect on lightoutput material that exhibits unacceptable lightoutput degradation. In fact, the 100°C test had a lower degradation in SuperFlux LEDs. Lumileds only lightoutput degradation than the tests run at uses batches of material with an average light cooler temperatures. This illustrates that drive output change of less than 35% after 1000 hours current has a larger impact than ambient of operation at 70 mA. Table 6.4 Operating Life Test Results for HPWT-MHOO LEDs Operating Life Tests Average Change in Lightoutput (after 1.000 hours) Condition A, HTOL, 100°C, 30 mA Condition B, HTOL, 85°C, 50 mA Condition C, HTOL, 55°C, 70 mA Condition D, RTOL, 25°C, 70 mA Condition E, LTOL, 40°C, 70 mA 10% 21% 24% 24% 16% 9 Table 6.5 WEATHERABILITY TESTS Weatherability Tests Average Change in Lightoutput (after 1.000 hours) Condition A, WHTOL, 85°C, 85% relative humidity, 50 mA 19% Figure 6.6 Change in light-output for HPWT-xHOO LEDs when operated in Arizona weathering tests, at 70mA and outdoor temperatures. Figure 6.7 Photograph of apparatus used for Arizona accelerated sun exposure tests. 10 Company Information Lumileds is a worldclass supplier of Light Emitting Diodes (LEDs) producing billions of LEDs annually. Lumileds is a fully integrated supplier, producing core LED material in all three base colors (Red, Green, Blue) and White. Lumileds has R&D development centers in San Jose, California and Best, The Netherlands. Production capabilities in San Jose, California and Malaysia. Lumileds is pioneering the highflux LED technology and bridging the gap between solid state LED technology and the lighting world. Lumileds is absolutely dedicated to bringing the best and brightest LED technology to enable new applications and markets in the Lighting world. LUMILEDS www.luxeon.com www.lumileds.com For technical assistance or the location of your nearest Lumileds sales office, call: 2002 Lumileds Lighting. All rights reserved. Lumileds Lighting is a joint venture between Agilent Technologies and Philips Lighting. Luxeon is a trademark of Lumileds Lighting, LLC. Product specifications are subject to change without notice. Publication No. AB206 (Sept2002) 11 Worldwide: +1 408-435-6044 US Toll free: 877-298-9455 Europe: +31 499 339 439 Asia: +65 6248 4759 Fax: 408-435-6855 Email us at [email protected] Lumileds Lighting, LLC 370 West Trimble Road San Jose, CA 95131