Internal modulation low light detection technology

1 Introduction
Light is an important carrier for carrying and transmitting information. Photoelectric signals, as an important carrier of information, have the characteristics of large information capacity, easy control, convenient long-distance transmission and online measurement. At present, the detection technology of weak light has become a very important research method in many fields such as life science, material science, environmental science, food science and aerospace science, and various instruments related to it have become extremely important equipment in various laboratories.
In the low-light detection system, photoelectric conversion is its core part. We have studied photoelectric detection devices for 20 years. On the basis of fully analyzing the basic functions of photoelectric detectors, we proposed the new concept of "indirect coupling photoelectric detection"〔1〕. The internally modulated photoelectric detector (referred to as internally modulated photoelectric tube or internally modulated photosensitive tube) developed according to the new concept of "indirect coupling photoelectric detection" can directly convert the incident light signal into a modulated AC signal output without a mechanical chopper. We use this internally modulated photoelectric detection device as the photoelectric conversion part of the low-light detection system, so that the detection system can obtain a modulated AC signal without a mechanical chopper, overcome the difficulty of amplifying weak DC signals, and lay a foundation for improving the signal-to-noise ratio and reliability of the system. Moreover, the working conditions of the internal modulation photodetector are much simpler than those of the photomultiplier tube, and its service life is relatively longer. Therefore, the low-light detection system using the internal modulation photodetector as the photoelectric conversion device represents a new low-light measurement technology〔2〕.
2 Principle of internal modulation photoelectric detection
After the weak light signal is converted into an electrical signal by the internal modulation photodetector, it must undergo various signal processing such as amplification and filtering〔3〕. In the process of detecting weak light, the photoelectric detection system will always be interfered by some useless signals when working. For example, the interference of random fluctuations of photoelectrons in photoelectric conversion, the influence of the channel and the
background light on the radiation field during transmission, the interference noise introduced by the amplifier, etc. The main purpose of photoelectric signal processing is to suppress noise to the maximum extent and extract the useful information carried by the signal. The noise of the photoelectric detection system is shown in Figure 1.

These noises mainly come from two aspects: (1) from the outside of the research system, usually caused by electrical, magnetic, mechanical and other factors. These interferences are mostly regular and can be reduced or eliminated by taking appropriate measures. (2) from the natural interference of materials, devices and inherent physical processes inside the system being studied. For example, thermal noise caused by the irregular motion of charged particles in any conductor, shot noise caused by photon counting during light detection. These processes are random processes, and their size and regularity cannot be accurately predicted and cannot be completely eliminated, but their statistical laws can be known and some measures can be taken to control them.
Intrinsic noises in photodetectors mainly include thermal noise, shot noise, generation-recombination noise (g-r noise), temperature noise, etc. Noise is extremely harmful in actual photodetection systems. Since noise is always mixed with useful signals, it affects the correct detection of signals, especially weak signals. The ultimate detection capability of a photodetection system is often limited by the noise of the detection system. How to reduce the impact of noise is an important issue in the detection system.
2.1 Principle of Maximum Signal-to-Noise Ratio
When the detected signal light is very weak, the signal-to-noise ratio (S/N) of the photoelectric signal obtained after conversion by the photodetector is very small, which requires some special weak signal detection methods to extract the signal from the noise〔4〕. In order to obtain the maximum signal-to-noise ratio from the signal processing system, the frequency function of the system and the input signal should satisfy a certain relationship. Assume that the signal processing system is shown in Figure 2. Its H (jω) is the frequency function of the system; h (t) is the impulse response of the system; Si (t) is the input signal; So (t) is the output signal; Wi (ω) is the power spectral density of the input white noise, and Wi (ω) = No; PN (t) is the output noise power; Wo (t) is the power spectral density of the output noise; Si (jω) is the input signal spectrum; So (jω) is the output signal spectrum.

According to the frequency domain analysis method of the signal, the spectrum of the output signal can be obtained as follows:

Fourier transform equation (1) to obtain the time domain expression of the output signal

According to Figure 2, the output noise power spectrum density is In

this way, the power signal-to-noise ratio of the system output at time td can be obtained as follows

Using the Schwartz inequality to simplify it, it is arranged as follows When it satisfies From the above analysis, it can be seen that in order to obtain the maximum output signal-to-noise ratio, the frequency response function of the signal processing system and the spectrum of the output signal need to satisfy equation (8). The signal processing system that satisfies this relationship is called a matched filter, and matched filtering technology is an important method for weak signal detection. 2.2 Cross-correlation detection principle Using the different characteristics of signal and noise in correlation characteristics is a common method for weak signal detection. Definition of cross-correlation function: It describes the degree of correlation between x (t) at time t and y (t) at time t-l. If the occurrence of two functions (processes) has nothing to do with each other (such as signal and random noise), their cross-correlation function is a constant, which is equal to the product of the average values ​​of the two functions. If the mean value of one function (such as noise) is zero, then their cross-correlation function is zero. If the two functions have the same fundamental frequency, then the cross-correlation function retains the amplitude and phase information of the original function.

The principle of cross-correlation detection is shown in Figure 3. Given the repetition period or frequency of the input signal S(t), a reference signal y(t) with the same repetition period as the input signal is correlated with the input signal mixed with noise n(t). The cross-correlation function

Rsy(l) contains the information carried by the signal S(t), so the signal S(t) can be detected.
3 Design and structure of analog circuits
The basic principle of the signal processing part of the internal modulation low-light detection system is shown in Figure 4, which is mainly composed of a signal channel, a reference pulse, a multiplier, a sampling and holding circuit, an automatic range shifting circuit, and a precision voltage stabilization circuit.
The AC current signal output by the internal modulation detector is converted into an AC voltage signal after I-V conversion, and its mathematical expression is Vi=Vm cosωt. Since the signal-to-noise ratio of the signal is very small, the frequency-selective amplifier circuit must improve the signal-to-noise ratio of the signal. According to the maximum signal-to-noise ratio principle, the frequency function of the frequency-selective amplifier circuit is:

From the above formula, we can know that the frequency function of the frequency-selective amplifier circuit is an impact pulse, which can be realized by a narrowband bandpass filter. We use a Butterworth second-order bandpass filter. The Butterworth second-order filter has the flattest frequency-amplitude characteristic in the passband, which can well improve the signal-to-noise ratio of the detector output signal. The structure of the second-order Butterworth filter is shown in Figure 5. The center angular frequency Through appropriate adjustment, its center angular frequency is the same as the frequency of the detector output signal, so that the amplitude gain of the detector output signal after passing through the filter is maximized. In this way, the peak value of the output signal of the frequency-selective amplifier is proportional to the light intensity.

From the experimental results, the internal modulation low-light detection system uses the internal modulation photoelectric characteristics of the detector to convert the optical signal into a modulated AC signal, which facilitates the processing of the subsequent circuit and overcomes the shortcomings of traditional weak light detection. This internal modulation low-light detection technology is a basic technology that can detect low light from ultraviolet to mid-infrared, and can also detect weak pulsed light. It has a wide range of uses in industry, agriculture, national defense, science and technology, environmental protection, medical care, food hygiene inspection and other fields. The internal modulation fiber colorimetric thermometer developed using this technology has been used for online temperature measurement in the metallurgical industry.