Technical Insight
Non-intrusive techniques help to predict the lifetime of LED lighting systems
New, rapid evaluation methods should help to bring researchers closer to defining the long-term performance of lighting systems based on LEDs, according to Jennifer Taylor of the LRC.
Purchasing a new lighting system is no small expense. Consumers therefore want to know what they re getting for their money, including an accurate prediction of how long the system will last. This is especially true when it comes to newer, evolving technologies such as LEDs, which have no comprehensive performance record in lighting systems. However, researchers at the Lighting Research Center (LRC) at Rensselaer Polytechnic Institute in Troy, NY, are getting closer to developing non-invasive methods for measuring the lifetimes of LED lighting systems. Such systems have become increasingly popular during the last five years and are becoming commonplace in traffic signals, exit signs and other forms of display lighting. LEDs are now displacing other types of lamps, particularly in applications requiring colored light, and in the future LEDs are poised to impact general illumination as well.
System lifetimesConsumers purchasing LED systems usually want to know how long the systems will perform so that they can estimate maintenance frequency and costs. Manufacturers typically estimate their systems lifespans to be the same as that of the individual LEDs used, often stating a lifetime of up to 100,000 h. However, according to Nadarajah Narendran, director of research at the LRC, this prediction is unlikely to be accurate.
"When LEDs are placed into a system, we have to take into consideration the packaging, the drive current, and the environmental conditions where the system is operating," he said. "All these factors affect the life of LEDs in a system." In other words, LEDs do not perform in the same way in a packaged system as they do under ideal laboratory conditions.
Based on this fact, Narendran and his students Eugene Hong and Yimin Gu completed two studies to define new methods of determining junction temperature in LED systems (Hong and Narendran 2003; Gu and Narendran 2003). Junction temperature is a primary factor in the degradation of LEDs, and is affected by the ambient temperature and the amount of power used to run a system. While methods for estimating junction temperature are not new, methods appropriate for systems have not been explored thoroughly until now.
"When we started this research, we knew that junction temperature had an effect on the degradation of light output from LEDs," said Narendran. "What we wanted to know was how we could determine the temperature at the junction without opening the system, and if we could use that information to predict system life without long-term life tests."
Narendran says that taking a system apart may affect its integrity and operating life, while long-term life tests are not practical because they are time-consuming, sometimes taking years to complete. Therefore, these studies focused on finding alternative methods for determining junction temperature, which could then potentially lead to predictions of LED system life.
Determining junction Two common methods for determining LED junction temperature are measuring the LED lead wire temperature and measuring the potential difference across the junction. However, both of these methods require access to the lead wire of the LED, which means disassembling the system.
Based on a number of previous studies regarding LED emission spectra and how they change with temperature, Narendran and his team chose to investigate whether spectral measurements, which are non-invasive, could be correlated to junction temperature. Because different semiconducting materials and device packaging affect light output degradation in different ways, the LRC researchers started their investigation with the most commonly used 5 mm red (AlGaInP) and white (GaN+YAG:cerium phosphor) LEDs.
The team s first step was to pick a known, reliable method for measuring the junction temperature against which to compare results from spectral measurements. Because the potential difference across the LED junction changes linearly as the temperature is increased within a certain range, junction temperature can be determined by measuring the potential difference across the LED junction. The next step was to compare the measured junction temperature against the spectral power distribution (SPD) of the LED emission.
System lifetimesConsumers purchasing LED systems usually want to know how long the systems will perform so that they can estimate maintenance frequency and costs. Manufacturers typically estimate their systems lifespans to be the same as that of the individual LEDs used, often stating a lifetime of up to 100,000 h. However, according to Nadarajah Narendran, director of research at the LRC, this prediction is unlikely to be accurate.
"When LEDs are placed into a system, we have to take into consideration the packaging, the drive current, and the environmental conditions where the system is operating," he said. "All these factors affect the life of LEDs in a system." In other words, LEDs do not perform in the same way in a packaged system as they do under ideal laboratory conditions.
Based on this fact, Narendran and his students Eugene Hong and Yimin Gu completed two studies to define new methods of determining junction temperature in LED systems (Hong and Narendran 2003; Gu and Narendran 2003). Junction temperature is a primary factor in the degradation of LEDs, and is affected by the ambient temperature and the amount of power used to run a system. While methods for estimating junction temperature are not new, methods appropriate for systems have not been explored thoroughly until now.
"When we started this research, we knew that junction temperature had an effect on the degradation of light output from LEDs," said Narendran. "What we wanted to know was how we could determine the temperature at the junction without opening the system, and if we could use that information to predict system life without long-term life tests."
Narendran says that taking a system apart may affect its integrity and operating life, while long-term life tests are not practical because they are time-consuming, sometimes taking years to complete. Therefore, these studies focused on finding alternative methods for determining junction temperature, which could then potentially lead to predictions of LED system life.
Determining junction Two common methods for determining LED junction temperature are measuring the LED lead wire temperature and measuring the potential difference across the junction. However, both of these methods require access to the lead wire of the LED, which means disassembling the system.
Based on a number of previous studies regarding LED emission spectra and how they change with temperature, Narendran and his team chose to investigate whether spectral measurements, which are non-invasive, could be correlated to junction temperature. Because different semiconducting materials and device packaging affect light output degradation in different ways, the LRC researchers started their investigation with the most commonly used 5 mm red (AlGaInP) and white (GaN+YAG:cerium phosphor) LEDs.
The team s first step was to pick a known, reliable method for measuring the junction temperature against which to compare results from spectral measurements. Because the potential difference across the LED junction changes linearly as the temperature is increased within a certain range, junction temperature can be determined by measuring the potential difference across the LED junction. The next step was to compare the measured junction temperature against the spectral power distribution (SPD) of the LED emission.
