What is the application of diodes in blood glucose monitoring equipment?

1, The technical essence of photodiodes: precise conversion from optical signals to electrical signals
The core function of a photodiode is to convert optical signals into electrical signals through the photoelectric effect of a PN junction. When light of a specific wavelength is irradiated onto the PN junction, photon energy excites valence band electrons to transition to the conduction band, forming electron hole pairs (photo generated carriers). Under the action of reverse bias, the directional movement of charge carriers generates photocurrent, and its intensity is linearly related to the incident light power. This process involves three key parameters:

Quantum efficiency: directly determines the photoelectric conversion efficiency. For example, InGaAs photodiodes can achieve a quantum efficiency of over 90% at a wavelength of 1310nm, significantly improving weak light detection capabilities.
Response time: determines the speed at which the device captures changes in blood glucose concentration. PIN type photodiodes shorten the carrier transit time to the picosecond level by optimizing the intrinsic layer thickness, meeting real-time monitoring requirements.
Dark current: affects the accuracy of low concentration detection. The Low Dark Current 0.3mm InGaAs PIN photodiode developed by Beijing Minguang Technology has a dark current of less than 0.1nA and performs well in detecting weak light signals.
Taking the non-invasive blood glucose detector as an example, it uses dual wavelength laser diodes of 1310nm and 1550nm to irradiate the skin, and the photodiode array receives the diffuse reflection light signal. By measuring the absorption differences of light at different wavelengths and combining them with partial least squares regression (PLSR) algorithm, the influence of interfering substances such as water and protein can be eliminated, achieving accurate calculation of blood glucose concentration.

2, Non invasive blood glucose monitoring: a technological revolution driven by diodes
Traditional blood glucose monitoring requires puncturing the skin for blood collection, which poses a risk of infection and cannot be continuously monitored. The breakthrough of diode technology makes non-invasive monitoring possible, and its core principles include:

Near infrared spectroscopy absorption method: Glucose has characteristic absorption peaks in the 750-1850nm wavelength range. By emitting specific wavelength light through DFB laser diodes, the absorption intensity of glucose in tissue fluid is detected by photodiodes. For example, the 1550nm DFB laser produced by Sichuan Tengguang has a built-in TEC temperature control module, with power stability better than ± 0.5%, ensuring long-term monitoring reliability.
Photoacoustic effect method: When laser irradiates the skin, glucose absorbs light energy to generate ultrasound waves. After the ultrasonic sensor captures the signal, the photodiode converts the change in light intensity into an electrical signal. The wearable device developed by Tsinghua University adopts a three wavelength laser diode array and processes three sets of data through DSP fusion, with a detection accuracy of ± 10mg/dL.
Optical rotation detection method: Using the optical rotation characteristics of glucose, the concentration is calculated by measuring the deflection angle of transmitted light. Organic light-emitting diodes (OLEDs) as light sources, combined with photodiode arrays, can achieve non-contact detection and are suitable for dynamic blood glucose monitoring.

3, Multi wavelength fusion detection: a key technology to enhance anti-interference ability
The composition of human tissue is complex, and the light absorption characteristics of substances such as water, protein, and fat are similar to glucose, which can easily cause cross interference. Multi wavelength fusion detection improves accuracy through the following strategies:

Optimization of wavelength selection: Experiments have shown that the combination of 750nm, 980nm, and 1310nm wavelengths can cover the main absorption peak of glucose while avoiding the strong absorption region of water (1450nm). For example, a certain model of blood glucose meter adopts a dual wavelength design of 750nm and 980nm, and eliminates background interference through differential algorithm, with a detection error of less than 15%.
Dynamic tuning technology: By controlling the current of the laser diode to tune within a range of 15nm, real-time capture of changes in glucose absorption peaks is achieved. The physical simulation system shows that dynamic tuning can increase detection sensitivity by 40%.
Chemometric modeling: Combining partial least squares regression (PLSR) or support vector machine (SVM) algorithms, establish a nonlinear model of light absorption intensity and blood glucose concentration. Clinical data shows that the predictive correlation coefficient (R ²) of the three wavelength fusion model is 0.92, significantly better than that of the single wavelength model (R ²=0.78).

4, Anti interference design: system engineering to ensure clinical reliability
Blood glucose monitoring equipment needs to cope with multiple challenges such as environmental light, electromagnetic interference, and device noise. Anti interference design needs to be optimized from both hardware and algorithm levels

Hardware Design:
Optical filtering: Install a narrowband filter in front of the photodiode to suppress interference from non target wavelength light. For example, the bandwidth of a 1310nm filter can be controlled within ± 10nm, and the transmittance is greater than 90%.
Electromagnetic shielding: Metal casing is used to encapsulate photodiodes, reducing 50Hz power frequency interference. Experiments have shown that shielding design can improve the signal-to-noise ratio (SNR) by 20dB.
Low noise amplification: A JFET input operational amplifier is used to construct a transimpedance amplification circuit, reducing the input noise voltage density to 0.5nV/√ Hz. For example, the total noise of the circuit of a certain model of blood glucose meter is less than 0.3mV, which meets the requirement of 12 bit AD conversion.
Algorithm optimization:
Wavelet denoising: Decompose the photocurrent signal using the db4 wavelet basis to filter out high-frequency noise. Clinical tests have shown that wavelet denoising can improve signal smoothness by 35%.
Adaptive filtering: Using LMS algorithm to dynamically adjust filter coefficients and suppress environmental light fluctuations in real time. For example, under a background light of 1000lux, adaptive filtering can reduce detection error by 50%.
Temperature compensation: Monitor the junction temperature of the photodiode through a thermistor and correct the dark current drift using a lookup table method. Experiments have shown that temperature compensation can stabilize the detection error within the range of 25 ℃ to 40 ℃ within ± 8mg/dL.