3. Quantum Mechanics: Why Electrons Stop Behaving Like Marbles

Up to now, electrons have been little charged marbles obeying Newton's laws and Maxwell's equations. That picture works fine at the scale of a wire or a transistor channel. It breaks down completely inside an atom, inside a transistor's gate oxide, or inside a flash memory cell. To understand why your laptop's RAM forgets, why flash memory wears out, why diodes have a 0.7 V drop, and why bandgaps even exist, you need a few ideas from quantum mechanics. We will keep this short and intuition-driven.

3.1 The catastrophe of classical physics

By the year 1900, physicists thought they had it all figured out. Maxwell's equations explained electromagnetism. Newtonian mechanics explained motion. Thermodynamics explained heat. There were just three nagging little experimental results that did not quite fit. Within twenty-five years, those three little results had detonated all of classical physics and replaced it with something weirder.

Problem 1: Black-body radiation. Heat any object — a stove burner, a star — and it glows, emitting light. Classical physics predicted that a hot object should emit infinitely much energy at high frequencies (the so-called "ultraviolet catastrophe"). Real objects do not. They emit a smooth, characteristic spectrum that peaks at a wavelength depending on temperature, and falls off at high frequencies. Max Planck, in 1900, fixed it by guessing — and "guess" is the right word, he hated that he had to do this — that energy could only be emitted in discrete chunks of size $E = hf$, where $h = 6.626 \times 10^{-34}$ J·s is what we now call Planck's constant, the smallest quantity in the universe. With this rule baked in, the math worked, and the predicted spectrum matched experiment exactly.

Problem 2: The photoelectric effect. Shine light on a metal, and it knocks electrons out — but only if the light is above a certain frequency. Below that, no electrons fly out, no matter how bright you make the light. Above it, even a very dim light triggers the effect. Classically this made no sense; a brighter light has more energy, so it should always do better. Einstein, in 1905, said: light is not a continuous wave but a stream of particles, each carrying energy $E = hf$. A high-frequency photon has enough energy per particle to kick an electron loose; a low-frequency photon does not, no matter how many you throw. He won the Nobel Prize for this — not for relativity.

Problem 3: Atomic spectra. Heat a gas of hydrogen and it does not glow with a smooth spectrum. It emits only specific colors — sharp spectral lines. Each element has its own fingerprint of lines. Classical physics could not explain why. Niels Bohr, in 1913, proposed that electrons in atoms can only occupy certain "allowed" orbits, and they can only emit or absorb light when jumping between orbits. The energy of the photon emitted equals the energy difference between the orbits, and $f = (E_2 - E_1)/h$ gives the line's frequency.

The pattern across all three: energy comes in discrete chunks, not continuous flows. We say energy is quantized. The new physics that emerged is called quantum mechanics.

3.2 Wave-particle duality: electrons are also waves

The young French physicist Louis de Broglie made the audacious leap in 1924: if light, which we knew was a wave, can also act like a particle, then maybe electrons, which we knew were particles, can also act like waves. He proposed that every particle has an associated wavelength

$\lambda = \frac{h}{p}$

where $p$ is the momentum. For a baseball, this wavelength is so absurdly small ($10^{-34}$ meters or so) that it is utterly undetectable. For an electron, with momentum maybe a billion-billion times smaller, the wavelength is around $10^{-10}$ meters — about the size of an atom.

This is the most important fact in modern electronics. At atomic length scales, electrons behave like waves. They diffract, they interfere with themselves, and — critically for transistors — they can spread through thin barriers by a process classical mechanics would forbid.

3.3 The wavefunction and the Schrödinger equation

Quantum mechanics replaces "the electron is at position $x$ moving with velocity $v$" with "the electron is described by a wavefunction $\psi(x, t)$ whose square gives the probability of finding it at $x$ at time $t$." The wavefunction evolves in time according to the Schrödinger equation:

$i\hbar \frac{\partial \psi}{\partial t} = -\frac{\hbar^2}{2m}\frac{\partial^2 \psi}{\partial x^2} + V(x)\psi.$

Do not panic about solving it. The relevant fact is that solving this equation for an electron bound to a nucleus produces a discrete set of allowed energies — Bohr's quantized orbits, but now derived from first principles instead of guessed. The lowest energy is the ground state; higher energies are excited states. An electron jumping down from an excited state to a lower one releases the energy difference as a photon — and atom-by-atom, this is where every photon you have ever seen ultimately comes from.

For a single atom in isolation, those allowed energies are sharp. Now bring two identical atoms close together. Each atom's energy levels split into two slightly different levels — one for the electron being closer to atom 1, one for atom 2. Bring three atoms together, three levels. Bring $10^{23}$ atoms together (a typical macroscopic chunk of solid silicon), and the levels merge into a continuous band of allowed energies. That is exactly what a band in solid-state physics is — the smeared-out remnant of the original atomic levels when a huge number of atoms come together to form a solid. We will pick up this thread in section 4.

3.4 Heisenberg's uncertainty principle: a fundamental fuzziness

Werner Heisenberg discovered that nature does not allow you to know certain pairs of quantities simultaneously to arbitrary precision. The most famous pair is position and momentum:

$\Delta x \cdot \Delta p \geq \hbar/2.$

The more sharply you pin down position, the less defined momentum becomes (and vice versa). This is not a measurement limitation — it is built into the structure of reality. Particles literally do not have a perfectly defined position and momentum at the same time.

For us, the practical consequence: at small enough scales, "this electron is at position X with velocity V" is meaningless. We must talk about probability distributions. And the smaller our device, the more this fuzziness intrudes. Modern transistors are 5 nm — about ten atoms across — and their behavior absolutely requires quantum-mechanical analysis to model accurately.

3.5 Quantum tunneling: the impossible-but-real phenomenon

Here is the headline quantum-mechanical effect for hardware engineers.

Classically, if an electron does not have enough energy to climb over a barrier, it does not get past. End of story. Bouncing rubber ball that cannot quite reach the top of the wall: it falls back.

Quantum mechanically, the wavefunction does not stop dead at the barrier. It enters the forbidden region, and although its amplitude decays exponentially with depth, if the barrier is thin enough, the wavefunction comes out the other side. When it does, the electron has a nonzero probability of being on the far side. The electron has tunneled through.

The tunneling probability falls off exponentially with the barrier's thickness. For a barrier 10 nanometers thick, the probability is essentially zero. For one 2 nm thick, it is small but nonzero. For one 1 nm thick — the gate oxide thickness in a modern transistor — it becomes a serious leakage current. Tunneling is the reason chips made on processes below 22 nm leak power even when "off."

Where tunneling shows up in your devices.

• Flash memory. A flash cell stores a bit by trapping (or not trapping) electrons in an isolated "floating gate." To program a cell, you slam it with high voltage, encouraging electrons to tunnel through the oxide into the floating gate. To erase, you tunnel them back out. Each cycle slightly damages the oxide. After about 10⁵–10⁶ cycles, the oxide fails and the cell is dead. This is why SSDs wear out.
• Flash retention. Stored electrons can also tunnel back out very slowly even at zero voltage. This is the mechanism by which flash chips eventually lose data over years of unpowered storage. JEDEC specs require ten years of data retention for new flash, falling to one year for end-of-life flash. Old USB sticks left in a drawer often forget.
• Tunnel diodes. A heavily-doped diode where tunneling dominates conduction. Used in microwave oscillators in the 1960s. Mostly historical, but a beautiful demonstration of the effect.
• Scanning Tunneling Microscope (STM). A sharp metal tip is brought to within an atom's distance of a sample. Electrons tunnel between tip and sample, producing a current. Move the tip across the sample and the tunneling current produces an atom-by-atom map. STMs were what gave humanity its first images of individual atoms.
• Voltage-glitching attacks. A brief voltage drop across a chip can cause the energy barriers in transistors to thin temporarily, allowing a flurry of unintended tunneling. Used by attackers to glitch a chip's authentication check at exactly the right moment. Counter-defenses: voltage monitors that detect drops and trigger reset.

3.6 Spin and the Pauli exclusion principle

One last quantum fact, because it is important for understanding why bands fill up the way they do. Every electron has a property called spin — like a tiny intrinsic angular momentum — which can take only two values: "up" or "down." There is no in-between, and there is no "stop spinning"; the electron always has spin.

The Pauli exclusion principle says: no two electrons in the same atom (or, more generally, in the same quantum state) can have all the same quantum numbers. Concretely: each energy level can hold at most two electrons, one spin-up and one spin-down. Any further electrons must go into the next energy level.

This is what builds up the periodic table. Hydrogen has one electron, in the lowest level. Helium has two — both in the lowest level (one spin up, one spin down). Lithium has three; the third has to go into the next level up. Carbon has six. Silicon has fourteen. The chemistry of every element is determined by where its electrons sit in these levels. And — crucially for us — the bandgap structure of silicon is determined by exactly this quantum mechanical accounting.