Research‎ > ‎

obic2

ABSTRACT

Efficient electric-luminescence (EL) is the major purpose and function of light emitting devices. Much engineering effort and technical breakthrough lie in the developing of new materials and structures for LEDs. Therefore, mapping the EL with high spatial resolution from the LEDs is an important method in characterizing various properties of the LEDs, for instance, in failure analysis. In this study, we are reporting a new development in EL mapping that allows functional imaging in probing the dynamical behavior of the LEDs. Traditionally, the response time of a semiconductor device is measured as the whole response of the device. For comparison, our method allows point-to-point characterization of a device’s active area with high spatial resolution through scanning microscopy.

未命名 - 4 未命名 - 3 200K

(a) LED chip (b) DC EL (c) RF EL

(a) The LED chip’s image in micrography. (b) The electroluminescence image is induced by laser scanned by FV-300. The dark area is the electrode and the bright area is the light-emitting region. (c) The demodulated electroluminescence signals from the two phase-shifted channels are color coded in green and red, respectively. The change in color reflects the difference in response speed that is revealed through the detected phase on the lock-in circuit.



PRINCIPLE 



PN

(e) electric-luminescence diagram


In the LEDs, photons will be generated when an electron-hole pair recombined radiatively. At forward bias, the high rate of recombination will result in the electroluminescence. This process is determined by and reflects carrier dynamics within the device. Radiative recombination may be induced by the following mechanisms, including photo-luminescence by the material absorption of photons, cathode-luminescence by the electron beam excitation, and electro-luminescence by the carrier injection. For electroluminescence, when the injected current is modulated, the corresponding Fourier expansion of the current is

......(1)


If lock-in detection is employed, only the first harmonic term will be selected while the constant, I0, and the higher harmonics terms, , can be ignored. The electroluminescence signal, i(t), is then the convolution of the injected current with the temporal response of the device, R(t)


where

 


at t>0. It is straightforward to carry out the calculation to obtain

where

is the phase shift and contains the information of response time. The response time , ,can then be easily derived by the formula,


Note that the response of the device is represented by R(t) with a single primary time constant that the analysis can be simplified. Experimentally the pertinent phase information, , is determined by the lock-in amplifier that produces outputs at two different phases.


EXPERIMENTAL SETUP 

A commercially available laser scanning module (FV-300, Olympus) is adapted for the EL mapping. The confocal mapping optics of the FV-300 allows point-to-point characterization of the device. The time-resolved measurement with frequency domain method is achieved with the use of a RF lock-in amplifier (SR844, Stanford Research). Its working range is from 25 KHz to 200 MHz. To synchronize the phase sensitive lock-in loop (with millisecond response time) with the rather rapid galvano-mirror based scanning imaging mechanism, special modification is made to the control
module of the scanning microscope to enable external trigger
未命名 - 1
(d) electric-luminescence configuration


APPLICATION 

The green and red image are the EL signal output from Lock-in Amplifier. The color coded by Olympus fluoview software.



This poster shows the effectiveness of time-resolved measurements in monitoring the electroluminescence signal from the LEDs. The observed time-delay is likely caused by the finite drift velocity of the carriers within the device, which levies the ultimate limits on the response of the devices.


This study shows the effectiveness of time-resolved measurements in monitoring the electroluminescence and OBIC signal from the LEDs. The observed time-delay is likely caused by the finite drift velocity of the carriers within the device, which levies the ultimate limits on the response of the devices.