I don’t know if CGM sensor technology has changed drastically or not. I poked around about 5 years ago to attempt to learn more about on how CGM sensors work. The link below is to a May 2014 Diabetes Forecast non-technical overview article.
The sensor starts with an enzyme, glucose oxidase, which converts glucose + oxygen to gluconic acid + hydrogen peroxide. The hydrogen peroxide then reacts with platinum inside the sensor, resulting in (I think) oxygen, hydrogen gas, and an electron or two. The electrons are the source of the current from the sensor.
Medtronic pumps at least as recently as my 630G reported this current from the sensor as the “ISIG” value. I’m pretty sure that acronym just stands for “current signal”. (“I” is a standard symbol for electrical current in the tech world).
The reason the sensor is calibrated after it is initially inserted is to allow time for the chemical process to stabilize (pretty much). Immediately after insertion the ISIG from my Enlites oscillated back and forth but then eventually stabilized.
At its simplest, calibration amounts to assuming your blood glucose is directly proportional to the current (ISIG) from the sensor. You divide the blood glucose value entered at calibration by the ISIG value reported by the sensor to get a conversion constant. After calibration, the sensor ISIG value is multiplied by that calibration constant to get a glucose value.
That’s the view from 30,000 feet as I understand it. As the technology has progressed, I’m sure there are a lot more nuanced details to it now.