Low Frequency Noise and Random Telegraph Signal in a Multiple Quantum Well Infrared Photodetector
- PDF / 407,439 Bytes
- 6 Pages / 416.88 x 635.4 pts Page_size
- 91 Downloads / 187 Views
This paper will discuss the effects random telegraph signal (RTS) has on the low frequency noise characteristics of the device. It will be shown that as RTS becomes noticeable the noise floor of the device is raised by between 2 and 4 orders of magnitude. After a certain temperature the noise increase lessens to factors of 2 or less. At the temperatures above the appearance of RTS, the noise remains high but increases slowly with temperature. This supports the theory that the RTS does not disappear at higher temperatures but that individual events begin to overlap until the time signal appears to be simply a noisy signal with no distinct RTS events. EXPERIMENT The device used in this study consists of 20 periods of three quantum wells each. The three quantum wells are InxGal.xAs doped n-type with Si and absorb photon wavelengths of 11.8 jim (x=0.1), 9.7 jim (x---O.15), and 8.2 jim (x=O.2). Well widths are 7.0 nm, 6.5 nrn, and 6.5 nm, respectively. Barriers consist of undoped A10.07GaO. 93As and are 45 nm wide. Total device length including contact regions is 4.8 gim. After gold bonding the device to a TO-12 can, it was placed on a cold finger for testing at low temperatures. The measuring setup consisted of two AC coupled low noise amplifiers in series, with an attenuator in between them, and a low pass filter connected to an HP 54645A Digital Oscilloscope for time data and an HP3582A Spectrum Analyzer for noise spectral data. RTS time data was collected with a 50 j.s resolution in 50 s windows. No time data was acquired if there was no discernible RTS. Noise spectral data was collected from DC to 25 kHz. Analysis of the noise spectral data was carried out for values between 4 Hz and I kHz due to limitations of the spectrum analyzer and the frequency response of the amplification/attenuation circuitry. RESULTS Random Telegraph Signal RTS is a phenomenon, when coupled to the outside circuit, that manifests as a voltage or current pulse. It is excess noise that can be discussed with generation-recombination terminology. In a device that has a small number of mobile carriers, addition or subtraction of a single carrier is observable at the output. It appears as a sudden change in voltage or current magnitude in the appropriate direction to describe emission or capture of a single electron. A pulse is obtained when emission follows capture (or capture follows emission) by some finite time period before another capture (or emission) event occurs. The signal is considered random since there appears to be no periodicity associated with the events. A capture (or emission) time is the time width of an observed pulse terminated by a capture (or emission) event. The distribution of capture (or emission) times is Poissonian in nature. RTS appeared under low bias (1.5 RA) and low temperature (between 77 K and 180 K) for the device under study. There appears to be a range of temperatures where individual RTS events are clearly discernible in the time data. Below that range it is unlikely that RTS events are occurring. Above tha
Data Loading...
