Test method for current noise of high-resistance device and medium material

Abstract

The invention discloses a test method for current noise of a high-resistance device and a medium material, comprising the following test steps: firstly, arousing the low-frequency current noise of a high-resistance sample by utilizing a sample adapter; then amplifying the sample low-frequency noise by utilizing a low-noise current amplifier; collecting amplified noise signals by utilizing a data acquisition card; calculating the power spectrum density S0(f) of the collected current noise signals of which the high-frequency parts are attenuated by the amplifier; obtaining the amplitude-frequency characteristic curve of the current amplifier by utilizing a locking amplifier, and calculating a normalized curve Q(f) according to the amplitude-frequency characteristic curve; reducing the attenuated power spectrum density S0(f) by utilizing the Q(f), thus obtaining a power spectrum density curve S(f) of a reduced band spread; and finally, carrying out data analysis on the S(f) to screen devices or research sample reliability. By utilizing the test method, the problems of narrow times-number transmission band and insufficient test data in the existing test method for high-resistance sample noise are solved; and the test method provided by the invention has the advantages of high efficiency, automated operation and high precision of reduced data.

Description

technical field [0001] The invention belongs to the technical field of electronic testing and measurement, and relates to a current noise testing method for high-resistance devices and dielectric materials, which is used for testing the current noise of high-resistance resistance devices or high-resistance materials within a wide frequency range. Background technique [0002] 1. Current noise of devices and materials [0003] When a DC bias voltage is applied to both ends of a device (such as a resistor, capacitor), the internal charge of the device will move in a directional manner to form a current. The average value of the current conforms to Ohm's law, but the instantaneous value of the current fluctuates randomly. This fluctuation is caused by various reasons such as the random fluctuation of carriers in the device, the random capture of carriers by defects in the device, the release process, and the scattering of carriers by grain boundary barriers. We call this insta...

AI Technical Summary

Problems solved by technology

However, in the case of a certain bandwidth gain product, a larger amplification factor will make the bandwidth of the signal very narrow, so that the high frequency cutoff frequency of the signal is usually less than 100Hz. Figure 2 The relationship between SR570 low noise current amplification and gain we can see this

When the sample resistance is higher, the amplification factor needs to be modulated to be larger, which makes the high-frequency cutoff frequency of the signal only a few Hz, so the high-frequency signal is severely attenuated

[0016] (2), multi-stage amplifier will increase the system noise

However, after actual verification, people found that this method has serious technical problems, and it cannot be applied in most actual test situations.

Because the noise floor of the current amplifier itself is inversely proportional to the amplification factor, such as image 3 As shown, so the second-stage voltage amplifier will greatly amplify the noise floor of the front-stage current amplifier, so that the weak current noise signal to be measured is annihilated by the background noise of the system, and finally the signal cannot be identified

[0018] (3) Incomplete power spectral density data will lead to in-depth research on sample reliability

[0019] In noise analysis, in-depth analysis of test results needs to obtain a complete undistorted power spectral density curve, that is, the frequency band of the curve must be wide enough. People can't get a complete curve, so that researchers can't know whether there is G-R noise at higher frequencies. In addition, because the corner frequency of some samples is relatively high, it can't be obtained from the data, so people can't curve fit the data. or parameter extraction and other in-depth analysis

[0020] From the above analysis, it can be seen that the existing high-resistance sample current noise test methods still have serious defects, and the data with sufficient bandwidth cannot be obtained or the obtained data cannot effectively distinguish the real sample current due to the large background noise of the system. These defects largely restrict the application effect of current noise testing methods for high-resistance devices or materials

Images