Characterization of Light Emitting Diodes (Leds) and Compact Fluorescent Lamps (Cfls) by UV-Visible Spectrophotometry
Total Page:16
File Type:pdf, Size:1020Kb
Characterization of Light Emitting Diodes (LEDs) and Compact Fluorescent Lamps (CFLs) by UV-Visible Spectrophotometry C. Mark Talbott, Ph.D., Product Manager – Spectroscopy, Robert H. Clifford, Ph.D., Industrial Business Unit Manager, Shimadzu Scientific Instruments, Columbia, MD, USA Introduction The acceptance and commercial utilization of Compact Fluorescent Lamps (CFLs) and, more recently, Light Emitting Diodes (LEDs) have grown significantly in the past five years. This change was fueled in part by the United States Congress 2007 legislative mandate to eliminate sales, and phase out the use, of conventional incandescent lighting beginning in 20121. This mandate has ultimately increased research into CFL and LED technologies. This paper demonstrates the use of a typical laboratory UV-Vis spectrophotometer, with no modification to the bench or software, to measure and characterize CFL and LED lamps. Using a customized accessory, spectral characteristics such as peak wavelength (λp), Full Width at Half Maximum (FWHM), centroid wavelength (λc), dominant wavelength, color, and color purity were readily determined for a variety of commercial CFLs and LEDs. Compact Fluorescent Lighting technology is the current main alternative to incandescent lighting. Compact Fluorescent Lamps (CFL) reportedly consume 75% less energy than incandescent counterparts.2 CFL lamps do contain a small quantity of mercury which may offer a potential future safety risk both for homes and in landfills. A typical UV-Visible spectrum from a CFL is shown in Figure 1. Figure 1: UV-Vis emission spectra of a typical compact fluorescent lamp (CFL) (violet) and a Mercury lamp (green). Light Emitting Diodes (LEDs) are a viable alternative to CFL technology and manufacturers are beginning to offer LED options as alternatives to CFL lighting. LEDs are semi-conductor devices with a p-n junction. When a forward voltage is biased across the p-n halves of the junction, photons of light are emitted. Because of the large refractive indices of the junction materials, the photons tend to leave the LED die at narrow angles between the junction. Typical LED design technology places the LED die on an anode aluminum reflector. The cathode is bonded to the top of the die with a gold wire and the complete package is encapsulated in a polymer that has appropriate optical properties. The end of the plastic encapsulation can be shaped to form a lens, adding directionality to the LED’s output.3 A typical yellow LED design is shown in Figure 2. The reflector, gold wire, and LED die can readily be seen. In addition, the LED was biased with a low forward voltage to demonstrate that the photon emission is from the area between the p-n junction. Another LED design that is typical for white LEDs is to imbed a blue LED die in a luminescent gel (Figure 3). When the blue LED die is forward biased, the blue photons stimulate the gel to fluoresce, providing broad spectrum white light. A UV-Vis scan of such an LED (Figure 4) clearly shows the 455 nm blue excitation peak and the broader white spectrum. Figure 2: Typical LED Architecture Figure 3: White LED Architecture Figure 4: UV-Vis Emission Spectrum of a White LED This article will demonstrate the use of the Shimadzu UV-1800 scanning spectrophotometer to characterize and measure CFL and LED properties. A mercury lamp calibration post is offered as an optional accessory for the UV-1800 spectrophotometer. An arm was attached to this post that allowed vertical height adjustment (Figure 5). To the arm was attached a clamping assembly for LEDs and a mirror to direct light from CFLs housed outside the lamp compartment into the entrance slit of the monochromator. No modifications or changes were made to the spectrophotometer or its lamps to perform the tests. In addition, Shimadzu’s UVProbe software and UVPC Color Analysis software provided the functionality required to acquire all spectra collected and presented here. Figure 5: UV-1800 Spectrophotometer with modified mercury lamp calibration post including the LED clamp and mirror. LED Bandwidths Figure 6 shows the spectral emission of a red LED emitting at 634 nm and a red diode laser emitting at 653 nm. The bandwidth of the LED (16 nm FWHM) is not as narrow as that of the laser (1.5 nm FWHM). However, even with the large bandwidths, LEDs are considered to be monochromatic sources. This can be evidenced by examining the position of the CIE x,y color coordinates for the individual LEDs. Accurate spectral acquisition and measurement of LEDs require a spectrophotometer with a monochromator capable of resolving these narrow bandwidth sources. The UV- 1800 spectrophotometer’s 1nm bandwidth is well within the resolution requirement for accurate LED spectral acquisitions. Figure 6: UV-Vis spectral emission scans of a red LED at 634 nm and a red diode laser at 653 nm. LEDs can be designed to emit a specific narrow-band wavelength or dominant color. The wavelength that an LED emits is related to the bandgap energy of the semiconductor materials used in manufacturing the p-n junction. The equation, = 1.24 / eV, relates the LED emission wavelength to the bandgap energy for a specific LED. For visible LEDs, this limits the bandgap energy to between 3.10 to 1.55 eV. An array of LED colors can be readily purchased from the commercial market. Emission spectra for these LEDs were acquired with the UV-1800 using a constant current power source to power the LEDs under test so that the constant forward current (If) was limited to 19.18 milliamps. The forward voltage (Vf) was allowed to vary as needed by the LED. The spectral acquisitions were acquired using a fast scan speed and a fixed sampling interval of 0.5. UVProbe software was operated in the energy mode and the silicon photodiode gain was set at 3. Figure 7 shows the spectral acquisitions for each of these commercial LEDs exhibiting a given dominant color. Table 1 shows reported and measured values for each LED. Characteristic LED information obtained from these scans was: Peak Wavelength - Wavelength at the maximum spectral band energy Center Wavelength - Wavelength at the center of the Full Width Half Max boundary Forward Voltage (Vf) - Voltage forward biased across the LED Power - Measured voltage times current Figure 7: Spectral acquisitions of commercial LEDs to include colors of: UV361, UV375, UV400, Blue, Teal, Aqua, Green, Yellow, Orange, Red, Deep Red, and IR. Reported Measured Center LED Wavelength Volts mWatts Wavelength Bandwidth UV361 361.000 4.54 87.1 362.915 13.13 UV375 375.000 3.422 65.6 374.45 9.84 UV400 400.000 3.364 64.5 391.455 10.95 Blue 470.000 3.037 58.2 460.48 18.86 Teal 490.000 3.059 58.7 490.23 25.42 Aqua 505.000 3.41 65.4 505.61 28.78 Table 1: Measured and Green 525.000 2.818 54.0 518.41 27.22 calculated values for the Spectral acquisitions of Yellow 590.000 1.972 37.8 593.435 13.77 commercial LEDs to include colors of: UV361, Orange 610.000 1.965 37.7 607.595 14.95 UV375, UV400, Blue, Teal, Aqua, Green, Yellow, Red 630.000 1.919 36.8 633.245 16.37 Orange, Red, and Deep DeepRed 660.000 1.831 35.1 653.075 21.27 Red. Measured Voltage versus Center Wavelength 700 y = -112.79x + 829.57 R² = 0.8511 600 500 400 300 200 Center Wavelength (nm) Wavelength Center 1.5 2.5 3.5 4.5 5.5 Measured forward voltage (volts) Figure 8: Measured LED voltage versus measured center wavelength. Figure 8 demonstrates the linear relationship expected between the center wavelength (color) observed and the forward voltage. The spectral scans and measurements demonstrate the value of being able to monitor this relationship when comparing LEDs, in design, and in manufacture. CIE Chromaticity Values With the variety of CFL coating phosphors available, CFL technology offers many color options that may be more suitable for various environments and work spaces. The ability to measure and quantify color values is paramount to the successful design and testing of new CFL phosphors. Figures 9 and 10 show in tabular and graphical format, the calculated x,y chromaticity values for spectra acquired from various commercial compact fluorescent lamps. Figure 9: Calculated CIE x,y chromaticity values for various commercial compact fluorescent lamps. Figure 10: Plot of the calculated commercial CFL on the x,y chromaticity coordinates system. Two important abilities for successful manufacturing, quality control, and comparison of LEDs is the measurement of dominant wavelength and purity. The dominant wavelength is defined4 as the point on the International Commission on Illumination (CIE) 1931 coordinates that is intersected by a line that is drawn from a theoretical illuminant “E” which has (CIE) coordinates located at the center (x=1/3, y=1/3) through the x,y coordinate values calculated from the LED spectra. Purity is the ratio of the distance from the illuminant “E” chromaticity coordinates to the LED calculated coordinates over the distance from the illuminant “E” chromaticity coordinates to the coordinates of the dominant wavelength. Illuminant “D65” was used in the Color Analysis software as it has CIE x,y coordinates of (0.31271,0.32902) and most closely matches those of illuminate “E”. Figure 11 shows the calculated CIE chromaticity and dominant wavelength values for the LED series tested. Figure 12 shows the plot of those values on the CIE chromaticity coordinate system. The plot shows that the LEDs on the red end of the spectrum lie on the border of the chromaticity plot and by definition would be expected to have a high purity value.