Nadarajah Narendran

Life of LED-based white light sources

Even though light-emitting diodes (LEDs) may have a very long life, poorly designed LED lighting systems can experience a short life. Because heat at the p-n-junction is one of the main factors that affect the life of the LED, by knowing the relationship between life and heat, LED system manufacturers can design and build long-lasting systems. In this study, several white LEDs from the same manufacturer were subjected to life tests at different ambient temperatures. The exponential decay of light output as a function of time provided a convenient method to rapidly estimate life by data extrapolation. The life of these LEDs decreases in an exponential manner with increasing temperature. In a second experiment, several high-power white LEDs from different manufacturers were life-tested under similar conditions. Results show that the different products have significantly different life values.

Impact of dimming white LEDs: chromaticity shifts due to different dimming methods

The goal of this study was to characterize the chromaticity shift that mixed-color and phosphor-converted white LED systems undergo when dimmed. As light-emitting diodes continue to rapidly evolve as a viable light source for lighting applications, their color shift while being dimmed should meet the current requirements of traditional lighting sources. Currently, LED system manufacturers commonly recommend pulse-width-modulation or PWM dimming schemes for operation of LED systems. PWM has the ability to achieve lower intensity levels and more linear control of light intensity compared to continuous current dimming methods. However, little data has been published on the effect dimming has on chromaticity shift of white LED lighting systems. The primary objective of this study was to quantify chromaticity shifts in mixed-color and phosphor-converted white LED systems due to continuous current dimming and pulse-width-modulation dimming schemes. In this study, the light output of the LED system was reduced from 100% to 3% by means of continuous current reduction or PWM methods using a PC white LED system and a mixed-color RGB LED system. Experimental results from this study showed that the PC white LED system exhibited very little chromaticity shift (less than a 4-step MacAdam ellipse) when the light level was changed from 100% to 3% using both dimming schemes. Compared to PC white LEDs, the mixed-color RGB LED system tested in this study showed very large chromaticity shifts in a similar dimming range using both dimming schemes. If a mixed-color RGB system is required, then some active feedback system control must be incorporated to obtain non-perceivable chromaticity shift. In this regard the chromaticity shift caused by the PWM method is easier to correct than the chromaticity shift caused by the continuous current dimming method.

A noncontact method for determining junction temperature of phosphor-converted white LEDs

The goal of this study was to develop a non-contact method for determining the junction temperature of phosphor-converted white LEDs as a first step toward determining the useful life of systems using white LEDs. System manufacturers generally quote the same life values for their lighting systems that the LED manufacturers estimate for a single LED. However, the life of an LED system can be much different compared with the life of an LED tested under ideal conditions because system packaging can affect system life. Heat at the pn-junction is one of the key factors that affect the degradation rate, and thus the useful life, of GaN-based white LEDs. The non-contact method described in this manuscript, combined with LED degradation rates, can be used to predict white-LED system life without affecting the integrity of the lighting system or submitting it to long-term life tests that are time-consuming. Different types of LED packages would have different degradation mechanisms. Therefore, as a first step this study considered only the 5mm epoxy encapsulated GaN+YAG Cerium phosphor white LED. The method investigated here explored whether the spectral power distribution (SPD) of the white LED could provide the necessary information to estimate LED junction temperature. Based on past studies that have shown that heat affects the radiant energy emitted by the InGaN blue LED and the YAG Cerium phosphor differently, the authors hypothesized that the ratio of the total radiant energy (W) to the radiant energy within the blue emission (B) would be proportional to the junction temperature. Experiments conducted in this study verified this hypothesis and showed that the junction temperature can be measured non-invasively through spectral measurements.

A method for projecting useful life of LED lighting systems

This study investigated a non-invasive method to determine the junction temperature of AlGaInP light-emitting diodes (LEDs) in a system. Because the primary cause for the AlGaInP LED degradation is junction temperature, this method can be used to predict LED life. Currently, life estimates of LED lighting systems quoted by manufacturers (commonly 100,000 hours) are based on the average life of a single LED measured under specific laboratory conditions. In reality, rates of degradation are much different for LEDs in a system than for those in a laboratory environment because the packaging and the environmental conditions in which the system operates can affect LED performance. Current practices for estimation require time-consuming life tests to accurately predict the life of LEDs. Therefore, a rapid estimation method for LED life is needed. Based on previous studies, the authors chose to focus on the measurement of junction temperature and its relationship to LED degradation. The primary objective of this study was to verify that wavelength shift could be used to estimate accurately the junction temperature of 5mm epoxy encapsulated AlGaInP LEDs. In this study, the junction temperature was increased by changing the drive current while holding the ambient temperature surrounding the LED constant, and by changing the surrounding temperature while holding the drive current steady. Experimental results from this study showed that for commercial LEDs, peak wavelengths shift proportionally to junction temperature regardless of how the temperature is created at the junction, and that this linear relationship could be used as a direct measure of the junction temperature. Because the primary cause for the degradation of AlGaInP LEDs is junction heat, the light output degradation rate of these types of LEDs can be predicted by measuring the spectral shift. Therefore, LED systems can be evaluated without disassembly in their intended application.

Performance characteristics of high-power light-emitting diodes

A laboratory experiment was conducted to investigate the performance characteristics of the currently available high-power LEDs under various drive conditions and ambient temperatures. Light output degradation and color shift properties as a function of time were measured for five types of commercial high-flux LEDs, namely, red, green, blue, and white from one manufacturer, and a different high-flux white LED package from a second manufacturer. The major difference between the two manufacturers’ products is that the first uses a single LED die per package, and the second uses multiple dies within its package. LED arrays were tested under normal drive current and ambient temperature, normal drive current and higher ambient temperature, and higher drive current and normal ambient temperature. Because each LED type has to operate at a particular ambient temperature, all were tested in specially designed individual life-test chambers. These test chambers had two functions: one, to keep the ambient temperature constant, and two, to act as light-integrating boxes for measuring light output parameters. Overall, the single-die green and white LED arrays showed very little light loss after 2,000 hours, even though the current and the ambient temperature were increased. However, the red LED array seemed to have a high degradation rate. The white LEDs had a significant color variation (of the order of a 12-step MacAdam ellipse) between them. However, the color shift over time was very small during the initial 2,000-hour period. For white LEDs to be accepted broadly for general illumination applications, the color variation between similar products must become much smaller, of the order of a 2-step MacAdam ellipse.

Spectral and luminous efficacy change of high-power LEDs under different dimming methods

Dimming is an important and necessary feature for light sources used in general lighting applications. An experimental study was conducted to quantify the spectral and luminous efficacy change of high-power colored and pc-white LEDs under continuous current reduction (CCR) and pulse-width modulation (PWM) dimming schemes. For InGaN-based blue, green, and pc-white LEDs, the peak wavelength shifts were in opposite directions for the two dimming schemes. The peak wavelength showed a blue shift with increased current, most likely due to band filling and QCSE dominated effects. InGaN LEDs exhibited red shifts with increased duty cycle, which is dominated by junction heat. AlInGaP red LEDs show mainly thermal-induced red shift with increased current or duty cycle. In addition, the luminous efficacy was always higher for the CCR dimming scheme at dimmed levels, irrespective of the LED type. Keywords: Light-emitting diodes (LEDs), white LEDs, mixed-color white LEDs, pulse-width modulation (PWM), continuous current reduction (CCR), peak wavelength shift, luminous efficacy

What is Useful Life for White Light LEDs

The goal of this paper is to initiate a discussion within the lighting community regarding standardized measurement procedures and a definition for useful life for light emitting diode (LED) technology. In general, LEDs do not fail catastrophically, but instead their light output slowly decreases over their operating period. Presently, some manufacturers use a 50% light output level as the criterion for LED life. Although 50% light loss might be acceptable for noncritical signage applications using monochromatic LEDs, it might not be acceptable for general lighting applications. It is important to develop a method for rating lamp life and a definition of “useful life” for LEDs so that when reported by manufacturers, the lighting community can compare LEDs to traditional light sources. The “useful life” definition for LEDs should consider light loss and color shift. Therefore, an experimental study was conducted to investigate light loss and color shift patterns of white LEDs as a function of operating time. The 5-mm type white InGaN +YAG LEDs evaluated in this experiment, representing technology commercially available in 1999, exhibited high light output degradation rates and color shifts as a function of operating time. It is further shown that using a simple mathematical fit to the data gathered during a short life-test study, and extrapolating it to predict the life of white LEDs, depends on the initial data collection period. Therefore, an alternate method for projecting LED life is investigated by overdriving the LEDs at different currents. Using their degradation patterns at higher drive currents, the life of these LEDs was predicted at normal drive current values. The results show excellent correlations between predicted light loss and actual measured losses at 20 and 30 mA drive currents for the LEDs tested. The authors believe that this technique is applicable for accurately predicting life of any type of LED and hope to verify this using future configurations. This study adds information to the knowledge needed for the lighting community to develop standardized measurement procedures and a definition for useful life for LED technology. INTRODUCTION Light emitting diodes (LEDs) were first developed over three decades ago. Most of the early LEDs were narrow wavelength band emitters with light output predominantly in the red to yellow region. During the 1990s Nakamura and colleagues (1994, 1997, 1998, 1999) demonstrated a blue LED based on gallium nitride (GaN). The development of the blue LED made the creation of the broad band white LED possible. Presently, white light is generated by combining the GaN-based blue LED and Y3Al5O12 (yttrium aluminum garnet or YAG) phosphor or by grouping red, green and blue LEDs in the correct proportions. [Refer to Stringfellow and Craford (1997) for an indepth discussion of LED technology.] The potential for significant energy savings and the potential for long life are the two major factors that have attracted this technology to the general lighting community. Over the past few years the technology has advanced significantly and some white LEDs presently available in the marketplace are rated at 10 to 15 lumens per watt (lm/W). White-light LEDs are among the first signs of an evolving solid-state technology for architectural lighting applications. Many industry experts are optimistic that solid-state technology will revolutionize the architectural lighting industry. Although luminous efficacies have been steadily growing for these LEDs, the amount of light generated by a single device is still low, usually under 1 lm. Manufacturers are actively working towards developing larger light output LED devices. Some of the methods presently used for achieving this goal include, grouping several smaller LED devices together, increasing the size of the semiconductor device, enabling the device to be driven at higher drive currents, and improving the light extraction efficiencies by shaping or modifying the emitting surfaces of the semiconductor device to prevent the light from being trapped within the cavity by total internal reflection (Ochiai-Holcomb et al., 2000; Windisch et al., 2000). The development of LED technology is fueled by the electronics industry, and as a result, the advances of this technology have been much faster than most lighting technologies. In anticipation of its widespread use for architectural lighting applications, many original equipment manufacturers (OEMs) have begun to develop light sources using white LEDs for the marketplace. Developing a new light source for general lighting and achieving high levels of application depends upon industrys success at standardizing product performance and at designing cost-effective products that reliably produce light of acceptable color. Other promising lighting technologies,

Improved Performance White LED

This paper describes work leading to the development of a new packaging method for white LEDs, called scattered photon extraction (SPE). Previous work by our group showed that the traditional placement of the phosphor close to the die negatively affects the overall luminous efficacy and lumen maintenance of phosphor-converted white LEDs. The new SPE method enables higher luminous efficacy by placing the phosphor at a remote location from the die and by shaping the lens surrounding the die to extract a significant portion of the back-transferred light before it is absorbed by packaging components. Although the remote phosphor concept is not new, SPE is the first method to demonstrate efficient extraction of back-transferred light and show over 60 percent improvement in light output and efficacy compared to similar commercial white LEDs. At low currents, the prototype white LEDs based on the SPE technique showed over 80 lumens per watt. The SPE concept was tried on two types of commercial packages and both showed similar improvements.

Characterization of thermal resistance coefficient of high-power LEDs

Heat at the junction of light-emitting diodes (LED) affects the overall performance of the LED in terms of light output, spectrum, and life. Usually it is difficult to measure junction temperature of a LED directly. There are several techniques for estimating LED junction temperature. One-dimensional heat transfer analysis is one of the most popular methods for estimating the junction temperature. However, this method requires accurate knowledge of the thermal resistance coefficient from the junction to the board or pin. An experimental study was conducted to investigate what factors affected the thermal resistance coefficient from the junction to the board of high-power LED. Results showed that the thermal resistance coefficient changed as a function of ambient temperature, power dissipation at the junction, the amount of heat sink attached to the LED, and the orientation of the LED with the heat sink. This creates a challenge for using onedimensional heat transfer analysis to estimate junction temperature of LEDs once incorporated into a lighting system. However, it was observed that junction temperature and board temperature maintains a linear relationship if the power dissipation at the junction is held constant.

Characterizing white LEDs for general illumination applications

During the past few years several manufacturers have introduced white Light Emitting Diodes (LEDs). At the present time these LEDs do not provide sufficient luminous flux for general lighting applications. Many manufacturers are studying the possibility of grouping several LEDs and overdriving them to produce more luminous flux. The impact of higher drive current on long-term performance of LEDs is not well known within the lighting community. Therefore, an experimental study was conducted to investigate the photometric characteristics of white LEDs as a function of time for different drive currents. The LEDs investigated in this study were the 5-millimeter type that uses GaN-based blue LEDs and Y3Al5O12 (yttrium aluminum garnet) phosphors (YAG phosphors). These LEDs produced 65 percent more light output at 55 mA compared to the light output at 20 mA. Groups of ten LEDs were driven continuously at constant current 20, 30, 50, 70, 90, and 110 mA and their relative light output were monitored at regular intervals for over 4000 hours. The light output degradation rate increased with increasing drive currents. Typically, LED manufacturers do not recommend driving these LEDs above 20 mA. However, it was noticed that the light output of these LEDs degraded to 65% of its initial value around 4000 hours even for those LEDs driven at 20 mA, which is the manufacturer recommended value for drive current. Considering the amount of flux produced by these 5 mm type white LEDs and their light output degradation rate, they are not yet suitable for general lighting applications.