NAND Flash Memory Explained | SuperSSD

A central part of any SSDs is where it actually stores data, which is typically on NAND Flash memory modules. There are other, similar types of memory modules – such as the now-discontinued Intel and Micron’s 3D Xpoint – but today NAND Flash completely dominates the market.

But what is NAND Flash memory, how is it made, and does it work? Let’s dive in.

What is NAND?

NAND is short for not-AND – a logic gate, which in turn is a device that performs a (Boolean) logical operation that results in a binary output. In the case of NAND Flash memory, the device in question consists of floating-gate MOSFETs, a type of transistor.

Compared to other types of computer memory such as DRAM, which loses data when power is turned off, NAND Flash is non-volatile, meaning that it doesn’t need power to maintain the information stored in it.  It is therefore suitable for long-term storage in devices like SSDs, USB drives, and memory cards.

There are four main types of NAND Flash memory: SLC (single-level cell), MLC (multi-level cell), TLC (triple-level cell), and QLC (quad-level cell). This refers to the number of bits that can be stored per memory cell, i.e. data density, which also affects endurance. SLC is the most expensive type with lowest density and highest endurance, whereas QLC is at the other end of the scale.

How is NAND Flash Memory made?

1. Basic Materials and Structure

• Silicon Wafers: The base material for NAND flash memory is silicon, typically fashioned into large, thin wafers. Silicon is chosen for its semiconductor properties, which are essential for creating electronic circuits.
• Memory Cells: The fundamental unit of NAND flash memory is the memory cell, which stores data. Each cell typically stores one bit of data.
• Floating Gate Transistors: Each memory cell is made up of a special type of transistor called a floating gate transistor. This transistor has two gates: the control gate and the floating gate.

2. Fabrication Process:

• Photolithography: The process begins with photolithography, where ultraviolet light is used to transfer a specific pattern (the circuit design) onto the silicon wafer. This pattern forms the blueprint for the memory cells and other components.
• Oxidation: The wafer is then subjected to oxidation, creating a thin layer of silicon dioxide on its surface. This layer acts as an insulator.
• Deposition and Etching: Various materials are deposited onto the wafer, and then etched away to create the desired structures. This includes forming the floating gate, control gate, and other components of the transistor.
• Doping: Certain areas of the wafer are doped with impurities to modify their electrical properties. Doping is used to create the n-type and p-type regions essential for the transistor’s functionality.

3. Creating the Floating Gate Transistor:

• Floating Gate: The floating gate is a key part of the NAND flash memory cell. It’s an electrically isolated gate that can hold an electrical charge. This charge determines whether the cell is in a ‘1’ or a ‘0’ state (binary data).
• Control Gate: The control gate sits above the floating gate, separated by a thin insulator. It controls the flow of electricity in the transistor and is used to read and write data to the floating gate.

3D NAND Flash Memory:

• In traditional 2D NAND, cells are arranged side by side in a flat, two-dimensional space. However, to increase storage density, manufacturers now focus on 3D NAND.
• Vertical Stacking: In 3D NAND, memory cells are stacked vertically in multiple layers. This allows for much higher density without increasing the physical size of the chip.

5. Testing and Cutting:

• After the memory cells are created and layered, the wafer undergoes testing to ensure functionality.
• It is then cut into individual chips, which are assembled into the final memory products like SSDs or memory cards.

Storing Data in the NAND Flash Memory Cell

Storing data in a NAND flash memory cell is a fascinating process that involves manipulating the electrical properties of the cell. Each cell in NAND flash memory typically stores one bit of data, although some advanced cells can store more. Here’s how it works:

1. Floating Gate Transistor: The core of a NAND flash memory cell is a floating gate transistor, which includes a control gate, a floating gate, source and drain regions, and an oxide layer separating the gates.
2. Electrical Charge: The key to data storage in NAND flash is the ability of the floating gate to hold an electrical charge.

Structure of a NAND Flash Memory Cell

Storing Data

1. Writing Data (Programming):

• Charging the Floating Gate: To store a ‘1’ or a ‘0’, an electrical charge is applied to the control gate.
• Tunneling: When a high voltage is applied, electrons are pushed through the oxide layer (a process known as tunneling) and trapped in the floating gate. This changes the threshold voltage of the cell.
• Charge Stored = 0: If the floating gate is charged, it blocks the flow of electrons from the source to the drain, which is interpreted as a ‘0’.

2. Reading Data:

• Detecting the Charge: To read the data, a lower voltage is applied to the control gate.
• Conductivity Check: The presence or absence of charge in the floating gate alters the conductivity between the source and drain.
• Interpreting Data: If the floating gate is charged (less conductive), the cell is read as a ‘0’. If it’s uncharged (more conductive), it’s read as a ‘1’.

Erasing Data:

• Removing the Charge: Erasing data in NAND flash is done at the block level, not for individual cells.
• Reverse Tunneling: A high negative voltage is applied, causing the electrons to tunnel back out of the floating gate, effectively resetting the cells in the block to a ‘1’ state.

Considerations

• Endurance: Each time a cell is programmed and erased, the oxide layer degrades slightly. This limits the number of write/erase cycles a cell can endure.
• Multi-Level Cells (MLC): Some NAND flash memory cells store more than one bit of data by holding multiple charge levels in the floating gate, enabling them to represent more than two states (e.g., 00, 01, 10, 11).

In summary, data storage in NAND flash memory involves altering the electrical charge of the floating gate in a transistor. This charge changes the transistor’s properties, allowing for the storage and retrieval of binary data. The precise control of these electrical charges and the ability to maintain them without power is what makes NAND flash memory so effective for digital storage.

Significance of 3D NAND Flash Layers

The term layer in SSD specifications, particularly when you see something like 112-layer TLC, refers to the number of layers of memory cells in a 3D NAND flash memory chip. This concept is part of 3D NAND technology, which is a significant advancement over traditional 2D (planar) NAND. Let me explain it further:

2D Vs 3D NAND

Initially, NAND flash memory cells were laid out in a single, flat layer. This is known as 2D or planar NAND. As demand for higher capacity increased, it became challenging to scale 2D NAND. Shrinking the cells to fit more on a chip led to issues like interference and reduced reliability.

3D NAND addresses these limitations by stacking memory cells vertically in layers. Instead of trying to fit more cells into a flat plane, cells are built upwards, like floors in a skyscraper. This allows for much greater storage density without needing to shrink each cell’s size, reducing interference issues and improving reliability.

The number of layers is a key factor in determining the storage capacity and performance of an SSD. For example, an SSD with 232-layer NAND will generally have higher storage capacity and potentially better performance compared to lower layer counts, all else being equal.