Tunneling Transistors

This article describes the principle of tunneling transistors and their advantages.

0040-77771 DPS ESC
The world of always-on PCs, tablets, and smartphones was born thanks to a remarkable trend: the increasing miniaturization of metal-oxide-semiconductor field-effect transistors (MOSFETs). MOSFETs, which are the basic building blocks of most integrated circuits, have shrunk to one-thousandth of their size in the past half century, from tens of microns in the 60s of the 20th century to only tens of nanometers today. As generations of MOSFETs become smaller and smaller, MOSFET-based chips run faster and are more power-efficient than ever before.
This trend has led to the longest and greatest series of victories in industrial history, giving us access to devices, capacity, and convenience unimaginable to previous generations. But this steady progress is under threat, and the heart of the problem lies in quantum mechanics. Electrons have a nerve-wracking ability to penetrate energy barriers – a phenomenon known as quantum tunneling. As chipmakers install more and more transistors on a chip, the transistors become smaller and smaller, so the distance between the different transistor regions is compressed. As a result, an electronic barrier that was once thick enough to block an electric current is now very thin, allowing electrons to pass through it quickly.

We have moved away from thinning the gate oxide, an important part of a transistor. This layer electronically separates the gate that controls the transistor turn-on and off from the conductive channel. By thinning this oxide layer, more charge can be channeled into the channel, speeding up the flow of current and allowing the transistor to run faster. However, the oxide thickness cannot be much smaller than 1 nanometer, which is what we can probably achieve today. Beyond this limit, there will be too much charge flowing through the channel when the transistor is in the "off" state, and ideally there will be no charge flowing at all. This is just one of several leaks.
We can't stop the electron tunnel from passing through this thin barrier, but we can make it work for us. In recent years, a newer transistor design – tunneling field-effect transistors (TFETs) – has accelerated. Unlike MOSFETs, which control the flow of current by raising or lowering the energy barrier, the energy barrier of a TFET remains high. The device controls turn-on and turn-off by changing the likelihood that electrons on one side of the barrier will appear on the other side.
This principle of operation is very different from the way traditional transistors work. However, this may be exactly what we need to do when MOSFETs stop evolving. It paved the way for the development of faster, denser, and more energy-efficient circuits to extend Moore's Law into the next decade.
This is not the first time that transistors have changed shape. Originally, semiconductor-based computers used circuits made of bipolar transistors. But just a few years after the introduction of silicon MOSFETs in 1960, engineers realized that they could make two complementary switches so that they could work together to form complementary metal-oxide-semiconductor (CMOS) circuits. Unlike bipolar transistor logic, this circuit only consumes energy when it is turned on. Since the first CMOS-based integrated circuits appeared in the early 70s, MOSFETs have dominated the market.
In many ways, MOSFETs are not much different from bipolar transistors. Both control the flow of electricity by raising or lowering the energy barrier – a bit like raising or lowering a sluice gate on a river. In this case, the "river water" is made up of two types of carriers: an electron and a hole, the latter being a positively charged entity that essentially lacks an electron from the outer shell of an atom in the material.
There are two permissible energy ranges, or bands, for these carriers. Electrons that have enough energy to flow freely in the material are located in the conduction band. Holes flow in low-energy bands, called valence bands, from one atom to another, much like an empty parking lot can become a full parking lot due to the constant flow of nearby cars in and out.
These bands are fixed, but we can change the energy associated with them by adding impurities or doping atoms to make the energy higher or lower, thus changing the conductivity of the semiconductor. n-type semiconductors doped with extra electrons conduct negatively charged electrons; P-type semiconductors that cause electron reduction through doping conduct positively charged holes.
If we combine these two semiconductor types, we get a misaligned band, creating a barrier in between. To fabricate a MOSFET, we inject a material between two complementary types, in n-p-n or p-n-p configurations. This creates three regions in the middle of the transistor: the source (where the charge enters the component), the channel, and the drain (the charge exit).

The two p-n junctions of each transistor provide an electronic energy barrier for charge flow, and the transistor can be turned on by applying a voltage to the gate above the channel. Applying a positive voltage to the n-channel MOSFET causes the channel to attract more electrons because it reduces the amount of energy required for electrons to move towards the channel. Applying a negative voltage to a p-channel MOSFET can have the same effect on the holes.
This simple way to lower the energy barrier is the most widely used current control mechanism in semiconductor electronics. Diodes, lasers, bipolar transistors, thyristors, and most field-effect transistors take advantage of this approach. However, there is a physical limitation to this approach: the transistor needs a certain amount of voltage before it can be turned on or off. This is because electrons and holes are always in motion due to thermal energy, and the most energetic part of them overflows the energy barrier. At room temperature, if the barrier is reduced by 60 millivolts, the current flowing through the barrier increases by a factor of 10; Each "decimal" current change requires a 60 millivolt change.

All of these current leaks occur below the threshold voltage of the device. The threshold voltage is the voltage required to turn on the transistor. Device physicists refer to this barrier reduction region as the subthreshold region, and a voltage of 60 millivolts per decimal is considered the minimum subthreshold swing. To keep energy consumption low, the subthreshold swing should be kept as low as possible. This reduces the voltage required to turn on the device, and the leakage current when turned off is reduced.
Sub threshold swings were not a big problem in the past, when chips needed higher voltages to operate. But now, subthreshold swings are starting to interfere with our efforts to reduce energy consumption. This is partly due to the fact that circuit designers want to make sure that their logic components have a clear distinction between the currents that define "0" and those that define "1". Transistors are typically designed in such a way that they can carry 10,000 times more current when they are on than they can leak when they are off. This means that to turn on a transistor, a voltage of at least 240 millivolts needs to be applied to it, i.e. 4 decimal currents, since 60 millivolts are required for each decimal.
In practice, CMOS circuits typically use a much higher operating voltage, close to 1 volt. This is because the most basic logic circuit in CMOS, the inverter, uses two series transistors. A NAND gate requires 3 series transistors, which means it requires a higher voltage than an inverter. If adjustments are to be made to account for process variability-which means that a wider voltage margin needs to be set to account for device-to-device variability-the voltage seen today is close to 1 volt to ensure operation.
These voltage requirements, combined with leakage issues, mean that MOSFET miniaturization is declining and there is no way out. If we want to further reduce the voltage to reduce energy consumption, there are two options (neither of which is attractive): we can reduce the current through the device, which reduces the start-up speed and thus sacrifices performance; Alternatively, the current can be kept high while allowing more current to leak out of the device at the time of shutdown. This is where TFET can be used. Unlike in MOSFETs, where the physical energy barrier between the source and drain is raised or lowered, in TFET, we use a gate to control the actual electrical thickness of the energy barrier, and thus the likelihood of electrons passing through the energy barrier.
Again, the magic of this approach lies in the p-n knot – but with some twists. In a TFET, the semiconductor material is housed in the configurations of p-i-n and n-i-p. where "i" stands for "intrinsic", meaning that the channel has as many electrons as the hole. The intrinsic state corresponds to the maximum resistivity possessed by a semiconductor. It also raises the energy associated with the bands within the channel, creating a thick energy barrier that charge carriers within the source are unlikely to cross. Both electrons and holes obey the laws of quantum mechanics, which means that their size is ambiguous. When the barrier is less than 10 nanometers thick, it is unlikely (but not entirely impossible) that electrons that are on one side of the barrier to begin on the other side.
In TFET, we increase this possibility by applying a voltage to the gate of the transistor. This overlaps the conduction band within the source and the valence band within the channel, opening a tunneling window. Note that in a TFET, electrons tunnel between the conduction and valence bands as they move to the channel. This is in stark contrast to what happens in MOSFETs. In a MOSFET, electrons or holes travel primarily through one band or another, all the way from the source through the channel to the drain.
Since the tunneling mechanism is not controlled by the flow of carriers across the energy barrier, the voltage swing required to start a TFET can be much smaller than that of a MOSFET. It is sufficient to apply enough voltage to make or move an overlap that makes the conduction band and valence band cross or not cross. (See illustration "Turn-off and on.") )