Diabetes is a metabolic disease characterized by hyperglycemia, and its main harm lies in various complications caused by long-term hyperglycemia [
1
]. Due to the increasing incidence of diabetes worldwide [
2
,
3
], continuous glucose monitoring (CGM) in the human body is becoming more and more important [
4
]. At present, the mainstream glucose sensors are divided into electrochemical sensors and optical sensors, among which electrochemical glucose sensors have attracted widespread attention for their low cost, rapid response and ease of use [
5
]. Electrochemical enzymatic reactions based on glucose and glucose oxidase or glucose dehydrogenase have been commercialized for blood glucose detection [
6
,
7
,
8
]. But these enzymes are easy to lose activity at temperatures above 40 °C and in acidic or alkaline environments [
9
,
10
], and immobilizing enzymes onto the surface of sensors is also a complex process [
11
]. Although many enzyme-free nanomaterials have been developed for the direct electrochemical oxidation of glucose [
12
,
13
,
14
], most of them cannot work under physiological pH and have poor selectivity [
15
]. These drawbacks greatly limit the long-term use of electrochemical glucose sensors in CGM systems. Optical glucose-sensing technologies, including fluorescence and surface plasmon resonance [
16
,
17
,
18
], detect the concentration of glucose (C
g
) based on changes in the properties of photons. They have the advantages of both sensitivity and versatility, enabling fast blood glucose monitoring [
19
,
20
]. However, fluorescent dyes suffer from the disadvantages of light bleaching and chemical instability, and most optical glucose sensors are large in size, expensive, have poor signal-to-noise ratio (SNR), and require frequent invasive calibration [
21
,
22
].
