Font Size: a A A

Study On The Dark Current And Relative Characteristics Of SW/MW Infrared Detectors

Posted on:2015-10-02Degree:DoctorType:Dissertation
Country:ChinaCandidate:H HuaFull Text:PDF
GTID:1228330422983176Subject:Microelectronics and Solid State Electronics
Abstract/Summary:PDF Full Text Request
Observing the first light, re-ionization, and galaxy assembly after the Big Bang isthe most heated topic among the researches of the astronomy, and the irradiation ofthe Big Bang is basically in the short-and middle-wavelength infrared band. HgCdTeis proverbially used to develop the SW/MW infrared focal plane arrays devices(IRFPAs) for space telescopes, because of it’s three qualifications which are low darkcurrent, big format and little bad-pixels.The dark current of the SW/MW HgCdTe IRFPAs becomes much smaller for theimprovement of the manufacture technology, which makes the measurement to beharder. On the other hand, the procedure of manufacturing HgCdTe IRFPAs ismulti-step and complex, so, many factors affect the bad-pixels rate of the devices.Meanwhile the format of the devices is becoming bigger, that makes the analysis ofthe bad-pixels more difficult.In this paper, two ways of testing the devices’ ultra-low dark current are studied,and the SW/MW HgCdTe IRFPAs’ dark current is tested and analyzed based ondevice processing approaches. Also, a special phenomenon called “responsivitybad-pixels pairs” is found out, and the reason is explained by looking into theabundant test data and picturing the3D indium appearance by X-ray CT technology.The primary coverage of this paper constitutes by four parts as below:1. The ultra-low current measurement method is studied and a new test systemis invented, which increases measurement accuracy by two or three orders. Thesources of system noise are made clear and reduced separately, then a system errorless than ten femto-ampere is achieved. A concept called vibration noise is stated andreduced by welding the test lead and gluing the bond point of the devices under test(DUT) especially. The possible photo current caused by cold shield (FOV=0)irradiance is calculated and confirmed to have little effect on dark current test results.The RC effect of the test bench on dark current measurement is discussed in detail,and the test source delay measure (SDM) cycle is optimized by experiments. 2. The multi-temperature dark current characteristics of SW/MW HgCdTeIRFPAs epitaxial grown on alternative substrates are measured and analyzed. Thetested RV curves of the MW devices manufactured by our own lab are fitted, whichindicates the dominant part of the dark current is G-R current at liquid nitrogentemperature and the descending sequence of the SRH lifetime is CZT-substrate,Si-substrate and GaAs-substrate. The multi-temperature (range from80K to200K)dark current characteristics of Si-substrate MW HgCdTe are obtained also, then twoor three orders lag are found between our devices with the foreign ones. Thedifferences in manufacturing processes of the devices from our lab to Sofradir arecompared, considering constitute of the dark current changes with temperature, tworeasons are figured out which depress the R0A of our devices, one is Hg vacanciesduring the formation of the p-type layer from data analysis of higher operatingtemperature, the other is ion implantation damage during the formation of the n-typeregion from data analysis of lower operating temperature.3. The dark current measurement method by long time integration is studiedalso. A practical plan is proposed, and SW/MW IRFPAs flipped with ROICs are testedby this method. Status, characteristics and difficulties of long time integration methodare researched in detail, a testing method named “similar FUR” sampling is stated,and devices with CTIA/DI ROIC are tested by this method separately. As a result, theglow of the ROIC is worked out, and the “like Reset Anomaly” appearance is foundand explained. Some proposals are proposed for improving the measurement ability oflong time integration method, which are adding a “FUR” sampling function for thetest bench, reducing the vibration noise of the cryogenic refrigerators, increase theperformance of the ROIC and etc.4. An analytical approach to research the bad-pixels of large format IRFPAs isdevised, and the formation mechanism of the “responsivity bad-pixels pairs” is foundout by this method. A method of classifying different type bad-pixels is stated. Themisleading effects of the testing software bugs and the improper long exposure timeset are researched. The “responsivity bad-pixels pairs” appearance is found and thereason is supposed to be caused by indium bump defects by a detail data analysis from the abundant test data of many of our devices. A device with “responsivitybad-pixels pairs” is cut to pieces and the flipped indium bump arrays in cut samplesare pictured by X-ray CT instrument, through which the corresponding defect indiumpairs are found out and it supports the indium bump detects opinion. Finally, a newway to automatically work out the height of the indium bumps is invented, which ismore objective and accurate comparing to the manual counting method.
Keywords/Search Tags:HgCdTe, SW/MW, dark current, long-time integral, bad-pixels
PDF Full Text Request
Related items