SESSION 1: Basic Storage Unit in Flash Memories

Most of us use flash memories today in n-number of devices, for example, SSDs, eMMC, SD cards, USB pen-drives, etc. The list is expanding day by day as we need a substitute for the mechanical and heat-plus-noise producing Hard Disk Drives (HDDs). 

I would like to specify that this article assumes a coneptual knowledge of MOSFETs and their working. If you are not familiar with it you can visit this article : Basics of MOS Devices

In this first session of Flash Memories, let us start from the unit cell of storage, that is, the particular electronic component that stores a bit of data. And that component is Floating Gate MOSFETs (FG-MOSFETs). Though I think I need not expand MOSFETs (Metal-Oxide-Semiconductor Field Effect Transistors), yet it would prove beneficial later!

So, this is what a traditional MOSFET looks like:

n-type MOSFET

The MOSFET has a METAL contact attached to the conducting polysilicon layer. Below the polysilicon layer is the insulating OXIDE layer followed by the SEMICONDUCTOR.

And this is how our FG-MOSFET looks like:

n-type FG-MOSFET

As clearly distinguishable from the image above, the FG-MOSFET has an additional oxide layer and a Floating Gate sandwiched between the two oxide layers. If somehow, a couple of electrons get trapped in this floating gate, ideally they won't be able to leak out even if the power to this device is turned off, thanks to the oxide layers on both sides. This electron trapping in the floating gate forms the basic concept of non-volatile storage using FG-MOSFETs.

So, for the very basic understanding, FG-MOSFETs can have one of the following two states:

1. Electrons are trapped in Floating Gate       : Programmed State, equivalent to BINARY STATE '0'.

2. Electrons are not present in Floating Gate  : Erased State, equivalent to BINARY STATE '1'.

Hence, if you are programming an FG-MOSFET, you are basically pumping electrons into the floating gate. And if you are erasing an FG-MOSFET, you are pulling out the electrons from the floating gate, obviously if the electrons are present. 

But the question now arises: How can you pump electrons into the floating gate with the insulating oxide layers surrounding it? Similarly how can you erase/remove the electrons from the floating gate? 

The answer lies in the keyword : tunneling.

Hence, if we apply a huge voltage (say 20 V) on the control gate and 0V (GND) at the substrate, the electrons present in the bulk in p-type semiconductor will tunnel into the oxide and get trapped in the floating gate. See the video below: 

Similarly, if we want to erase the device, we just reverse the voltages, that is, 20 V at substrate and GND on Control Gate. See the video below:

I guess the term tunneling must be familiar with you, but let us discuss this concept in a bit more detail in the next section. After that, we can be clear about program and erase in an FG-MOSFET!

Till then, "be constantly amazed with electrons!"