01 4Detailed Explanation of T Sensors

Currently, the sensors commonly used are typically 4T sensors, with their internal structure shown in the figure below:

Let’s start with a brief overview of the structure.

a PPD and FD for storing electronics

Before explaining the 4T, it’s necessary to examine two storage locations for electrons—PPD and FD—as shown in the figure below.

i PIN photodiodePPD

First, let’s look at the PIN photodiode, or PPD (Pinned Photodiode).

It will be used in the Sensor.Implantable Photodiode BPD（It uses a buried photodiode to detect light, enabling the conversion of incident external light into electrical signals:

Meanwhile, PPD is a special case of BPD, with an additional grounded P+ layer added on the N‑type region (the p++ layer highlighted in the red box in the figure above).

This design offers advantages such as ensuring complete electron transfer without leaving afterimages and suppressing dark noise (as shown in the figure below, image source: online), and has gradually become the mainstream approach.

During exposure, photons generate a photoelectric reaction on the PPD, converting into electrons and being stored in the N-well within the PPD (the n-well shown in the figure above).

ii Floating Diffusion Amplifier FD

In simple terms, the FD is a region where charge is converted into voltage.

During exposure, the accumulated charge is transferred to the FD, where it is converted into a voltage. The more photons that strike the sensor during exposure, the more electrons are generated, and the higher the resulting voltage will be. In other words, by reading the voltage at this point, one can determine the intensity of the incoming light.

b 4T: Four-transistor structure

Next, let’s briefly discuss transistors—here, “4T” refers to four transistors, namely the four positions shown:

Let me briefly explain what each of these four transistors does.

i Transfer gate TX/TG

In the preceding explanation, we noted that the electrons generated by PPD photosensitivity are stored in the PPD’s N-well, while the charge-to-voltage conversion takes place at the FD.

And here, the transfer gate (TX/TG) serves as a valve that controls the transfer of charge.

After exposure, when it is necessary to acquire light intensity information, the transfer gate is activated to move the charge to the FD node, where it is converted into a voltage. By reading this voltage value, the magnitude of the light intensity can be determined.

ii Reset RST

As the name suggests, RST stands for Reset.

Simply put, when we need to clear the electrons stored in the FD—sometimes clearing the PPD as well—we must trigger the RST operation.

If the cleaning here is not thorough, the next measurement will be affected.

To put it another way, when using a graduated cylinder to measure a liquid, you must empty the container completely before taking a reading; otherwise, the previous measurement will affect the current one.

iii Source Follower SENSE/SF

As mentioned above, the function of the floating diffusion amplifier (FDA) is to convert charge into voltage.

However, the voltage here has little driving capability, so it needs to be passively buffered by an SF stage.This voltage variation is reproduced exactly and fed to the subsequent readout circuit for signal readout..

To put it another way, FD is like a cup of water with a certain height.

The purpose of the SF is that, regardless of the type of container provided at the backend, it will draw water from the tap so that the water level in the container matches that in the cup, without affecting the original volume of water in the cup.

iv Row Select (RS)

Finally, for the remaining columns, select RS.

This one is fairly straightforward from its name—this switch reads the information of a specific column.

Let’s take a look at this schematic of the CMOS; the row select is represented here by RS:

As you can see, the CMOS sensor reads out one row of data at a time.

When we need to read out the first row of data, we activate the RS signal for that row and sequentially sample the SF voltage values of each pixel in the row, thereby completing one row of exposure.

Afterward, the RS of this row is turned off, the RS of the next row is turned on, and the data is read out sequentially, thereby achieving line-by-line exposure.

This is CMOS’s progressive scan exposure—the jelly effect.

02 CMOS with other structures

Next, we’ll discuss CMOS with a 3T structure.

a 3T CMOS

This discusses early 3T CMOS, which is not the same as a JFET sensor.

Let’s start by discussing the early 3T CMOS.

As the name suggests, 3T CMOS uses one fewer transistor than 4T CMOS, as shown in the first row of the figure below:

As we can see, the 3T CMOS circuit contains only three transistors: a reset transistor, a source follower, and a row select transistor.

When reading light intensity information, the source follower in a 3T CMOS sensor directly reads the charge stored in the photodiode. While this design saves one transistor, it cannot implement correlated double sampling (CDS), resulting in significantly higher noise, dark current, and afterimage artifacts compared to a 4T CMOS sensor.

Meanwhile, the PPD–FD separation, full transfer, and complete discharge features offered by 4T CMOS are capabilities that 3T CMOS cannot achieve.

b 5T, 6T, or even higher resolution CMOS

In fact, there is relatively little information available on 5T, 6T, and even higher node CMOS technologies.

However, the common view is that as the number of transistors increases, the sensor’s capabilities also expand.

For example, functions such as global reset and global shutter (charge storage) can be implemented.

Of course, the more transistors a CMOS device has, the more powerful its functionality becomes, but this also requires more advanced manufacturing processes and higher costs.