Henry Joseph Round, an English engineer, was examining the rectifying current behavior of SiC crystallites in 1907 and observed a faint yellow light was emitted from the SiC. This was both the first successful operation of a light emitting diode (LED) and the first reported recording of the phenomena of electroluminescence.1,2
Round’s work was expanded on by a few others in the early 20th century,2,3 but it would take decades for LEDs to become good enough for actual use.2 In the 1960s the first commercial LEDs were produced which had emission in the red and NIR spectrum regions.
It would take a further 30 years, in the late 1990s, before the breakthrough of high efficiency InGaN based green and blue LEDs. Modern LEDs are amongst the most versatile and commonly used light sources with an ever popular variety of applications. They are bright, efficient, and reliable. LED technology is advancing quickly, with new doping methods and device architectures being made in order to increase LED efficiencies and brightness.
There are a number of new disruptive LED technologies that are set to move LED design away from the classic type III-V semiconductors that have previously dominated the technology:
- Quantum dots (QDLEDs)
- Organic semiconductors (OLEDs)
- Halide perovskites (PLEDs)
An important method to characterize new LED designs and advance their development is electroluminescence (EL) spectroscopy. In EL spectroscopy a current is passed through a light emitting device to study the characteristics of the emitted light, in a time-resolved measurement or steady state.
A constant current is passed through the device and the EL emission spectrum calculated with a monochromator in steady state EL spectroscopy. The color rendering index and chromaticity coordinates of the emission can be calculated from the emission spectrum, as can the fundamental material properties like the bandgap of the semiconductor.
Using time-resolved EL spectroscopy the EL reaction of the device to short voltage pulses is observed in order to explore the dynamics of charge carriers in the device, for example triplet formation in OLEDs.
Figure 1. FS5 Spectrofluorometer which can be equipped with a range of source measure units and function generators for the measurement of steady state and time-resolved electroluminescence.
In this instance the FS5 Spectrofluorometer is equipped with the electroluminescence accessory and employed to analyze the emission characteristics of four type III-V LEDs and establish their chromaticity coordinates and bandgaps. An LED is made up of layers of semiconducting material that are either p-type or n-type.
An n-type semiconductor has an excess of electrons, a p-type has an excess of holes. The first commercial LEDs paired a p-type and n-type region to create a simple p-n junction structure where electrons and holes recombined at the interface to produce light. This simple structure is not very efficient, so to maximize LED efficiency, modern LEDs utilize a double heterojunction structure.
Figure 2 shows a diagram of a double heterojunction structure, consisting of two p-n junctions at the interfaces between three different semiconductor regions. When a voltage is applied across the semiconductor the electrons in the n-type region and the holes in the p-type region will drift nearer the p-n junctions and enter the active region.
The holes and electrons become energetically trapped in this region and consequently are more likely to recombine, since the active region has a lower bandgap than the flanking n-type and p-type regions. The recombination of the holes and electrons creates light with an energy equal to the bandgap (Eg) of the active region.
Figure 2. Double heterojunction light emitting diode band structure. Holes are shown as hollow circles and electrons as filled circles. Recombination of electrons and holes generates photons with an energy equal to Eg (active). Adapted from Wagner.4
Most conventional LEDs are made up of III-V compound semiconductors. A III-V semiconductor is an alloy containing elements from groups III (B, Al, Ga, In) and V (N, P, As, Sb) of the periodic table. The active region of the LED is formed from InxGa1-xN, where the inclusion of indium lowers the bandgap of the semiconductor.
The bandgap and consequently the emission wavelength and color of the LED can be tuned by adjusting the ratio of In to Ga. In, Ga and N are the elements used for green and blue LEDs. The n-type and p-type regions in these LEDs are constructed from GaN which has a large bandgap of 3.4 V (360 nm), and the ratio of Ga to N is altered to make it n-type or p-type.
InGaN LEDs become more inefficient and so it is not practical to make LEDs with peak emission wavelengths greater than ~550 nm utilizing InGaN as the In volume is increased. The most favored semiconductor to cover wavelengths beyond ~550 nm is (AlxGa1-x)0.5In0.5P.
By heightening the ratio of Al to Ga (higher x) the bandgap is increased and the emission wavelength of the LED is blue shifted. The peak emission wavelength of AlGaInP and InGaN LEDs can be calculated and the bandgap energy of the elemental composition and active region of the LEDs established by using electroluminescence spectroscopy.
Figure 3. Electroluminescence spectra of InGaN and AlGaInP LEDs showing the peak wavelengths and FWHM of the emission. The voltage applied to each LED was adjusted to achieve a current of 20 mA. Δλem = 0.1 nm.
Methods and Materials
Electroluminescence spectra were measured using the FS5 Spectrofluorometer equipped with a PMT-900 detector. Commercial light emitting diodes were purchased from Farnell Ltd; red (OVL-5526), blue (OVL-5523), orange (OVL5528), and green (OVL-5524). A source measure unit was employed to apply a voltage across the LEDs, with the voltage adjusted until the current through the LED was 20 mA.
Using the FS5, the emission spectra of four III-V LEDs was determined and is shown in Figure 3. The yellow and red LEDs are based on AlGaInP with wavelengths of 594 nm and 630 nm respectively, while the blue and green are based on InGaN and have peak emission wavelengths of 462 nm and 516 nm.
Generally, the emission of an LED is classified by its chromaticity coordinates instead of the peak wavelength for display applications. The chromaticity coordinates of the four LEDs were measured in CIE 1931 color space and are shown in Figure 4. The Fluoracle® software of the FS5 has a built-in wizard to produce a chromaticity plot from any emission spectrum, in either CIE 1976 or CIE 1931 color space.
Figure 4. Chromaticity coordinates of the emission from the four LEDs in CIE 1931 color space, calculated using Fluoracle from the electroluminescence spectra in Figure 3.
The bandgap of the active region of each LED can be easily measured by converting the peak wavelength of the LED emission into energy units (Table 1), using the information gained from the electroluminescence spectra. The bandgap energy can then be utilized to determine the composition of the semiconductor in the active region of the LED.
For AlGaInP the following relationship between the bandgap and the composition of the semiconductor has been found,2
|Eg = EgGaN + (EgInN – EgGaN) x – x (1–x) Eb
Where Eg is the bandgap of (AlxGa1-x)0.5In0.5P, Eg GaInP is the bandgap of Ga0.5In0.5P and x is the ratio of Al to Ga. Inserting the literature value of Eg GaInP of 1.91 eV,2 and the bandgaps from Table 1 into Eq. 2 gives x values of 0.30 and 0.10 for the yellow and the red LEDs respectively.
For InGaN LEDs an empirical relationship between the bandgap energy and the composition has previously been established to be:2
Where Eg is the bandgap of InxGa1-xN, Eg GaN and Eg InN are the bandgaps of GaN and InN which are 3.42 eV and 0.77 eV respectively,2 x is the ratio of In to Ga and Eb is the bowing parameter which has been empirically found to be 2.4 eV.2,5 Inserting these values and the bandgaps from Table 1 into Eq. 1 and solving the quadratic equation, gives x values of 0.16 and 0.23 for the blue and green LED respectively.
Table 1. Peak wavelengths, bandgaps and elemental composition of four III-V LEDs
||Peak Wavelength (nm)
The peak wavelengths, FWHM and chromaticity coordinates of the LED emission were measured from the electroluminescence spectra. The emission properties of four III-V LEDs were calculated using the FS5 Spectrofluorometer equipped with the electroluminescence accessory.
The bandgaps and elemental composition of the semiconductors were established from the peak emission wavelengths, showing the utility of electroluminescence spectroscopy. The capability of both the FS5 and FLS1000 spectrometers can be increased with a variety of electroluminescence accessories for the characterization of electrical devices.
References and Further Reading
- H. J. Round, A note on Carborundum, Electrical World, 49, 309 (1907)
- E. F. Schubert, Light-Emitting Diodes, 2nd ed., Cambridge University Press, Cambridge, NY, (2006)
- O. V. Lossev, Luminous Carborundum Detector and Detection Effect and Oscillations with Crystals, Philosophical Magazine 6 1024 (1928)
- E. P. Wagner Investigating Bandgap Energies, Materials, and Design of Light-Emitting Diodes, J. Chem. Educ. 93 1289 (2006)  L. Siozade, J. Leymarie, P. Disseix, A. Vasson, M. Mihailovic, N. Grandjean, M. Leroux, J. Massies, Modelling of Thermally Detected Optical Absorption and Luminescence of (In,Ga)N/ GaN Heterostructures, Solid State Commun. 115 575 (2000)
This information has been sourced, reviewed and adapted from materials provided by Edinburgh Instruments.
For more information on this source, please visit Edinburgh Instruments.