How Floating-Gate Transistors Trap Electrons

Flash memory relies on floating-gate transistors (FGMOS) to store digital data non-voluntarily, retaining information even when power is disconnected. By using quantum tunneling and high electric fields, these specialized semiconductor devices trap or release negative electrical charges (electrons) on an electrically isolated gate. The presence or absence of these trapped electrons modifies the voltage required to activate the transistor, enabling the cell to represent the binary digits—0s and 1s—that form the foundation of computing.

The Structure of a Floating-Gate Transistor

A standard floating-gate transistor resembles a traditional MOSFET (Metal-Oxide-Semiconductor Field-Effect Transistor) but includes an additional component. It consists of:

• Control Gate: The top gate connected to the memory circuit’s word lines, used to apply operating voltages.
• Floating Gate: A conductive polysilicon layer positioned beneath the control gate, entirely encapsulated by high-quality insulating dielectric layers (typically silicon dioxide). Because it is completely surrounded by insulators, any electrons placed on this gate remain trapped for years.
• Channel, Source, and Drain: The semiconductor pathway beneath the floating gate where electrical current flows when the transistor is turned on.

The Trapping Mechanism: Tunneling and Injection

To place electrons onto the isolated floating gate—a process known as programming or writing—electrons must overcome or pass through the insulating oxide barrier. Flash memory accomplishes this using two primary physical phenomena:

1. Fowler-Nordheim (F-N) Tunneling: A strong electric field is applied across the thin oxide layer by placing a high positive voltage on the control gate relative to the silicon substrate. This thins the energy barrier enough for electrons to physically “tunnel” through the insulator via quantum mechanical effects, entering the floating gate.
2. Channel Hot-Electron (CHE) Injection: A high voltage is applied across both the drain and the control gate. This accelerates electrons flowing through the channel to high kinetic energies (“hot electrons”), allowing them to overcome the insulator’s electrostatic barrier and jump directly into the floating gate.

Once inside the floating gate, the electrons are permanently trapped because the insulating oxide prevents them from discharging under normal operating conditions.

Translating Trapped Charges to Binary States

The trapped electrons alter the electrical properties of the transistor by modifying its threshold voltage (\(V_{th}\))—the minimum voltage required on the control gate to allow electrical current to flow from the source to the drain.

• High Trapped Charge (Programmed State / Binary 0): When electrons are trapped on the floating gate, their negative charge partially cancels out (screens) any positive voltage applied to the control gate. Consequently, a much higher voltage is required to turn the transistor on.
• No Trapped Charge (Erased State / Binary 1): In an uncharged floating gate, there is no negative screening effect, meaning a relatively low voltage on the control gate is sufficient to open the channel and allow current to flow.

Reading and Erasing Data

To read the stored bit, a reference voltage (intermediate between the high and low threshold voltages) is applied to the control gate: * If current flows through the channel, the floating gate has no trapped charge, registering as a 1. * If current is blocked because the threshold voltage was not reached, trapped electrons are present, registering as a 0.

To erase the cell and reset it to a binary 1, a reverse electric field is applied (typically a large negative voltage on the control gate or a high positive voltage on the substrate). This forces the trapped electrons to tunnel back out of the floating gate via Fowler-Nordheim tunneling, clearing the charge and returning the transistor to its default state.