Where color actually comes from
In the last guide we sorted materials by what they do to light. A metal is opaque and shiny because its sea of free electrons can soak up a photon of any visible energy and instantly spit it back out — so nothing gets through and everything bounces off. An ordinary window glass is clear because its electrons are locked in bonds with a band gap far wider than any visible photon, so light simply has nothing to grab onto and sails straight through. Color lives in the interesting middle: a material that absorbs some visible wavelengths but not others.
Here is the one idea that unlocks nearly all color. When a material selectively absorbs certain wavelengths from white light, the color you see is whatever is left over — the leftovers that got transmitted (if it is see-through) or reflected (if it is opaque). This trips people up constantly: a ruby is red not because it adds red, but because it swallows the green and violet and lets only red escape. The color you see is the complement of what got eaten. So the whole question of color reduces to a sharper one: which photon energies does this material absorb, and why?
For a semiconductor the answer is beautifully clean: it absorbs any photon energetic enough to kick an electron across its band gap, and transmits the rest. A photon's energy in electron-volts is just E = 1240 / (wavelength in nanometres), and the visible band runs from about 1.8 eV (deep red, 700 nm) up to 3.1 eV (violet, 400 nm). Now the numbers tell the story. Diamond's gap is a huge 5.5 eV — far above every visible photon — so it absorbs none of them and is water-clear. Silicon's gap is a mere 1.1 eV, below the whole visible range, so it absorbs every color and looks opaque dark grey. Cadmium sulfide sits right in the sweet spot at about 2.4 eV: it swallows everything bluer than 1240/2.4 ≈ 517 nm (violet, blue, green) and passes the yellow-orange-red, which is exactly why 'cadmium yellow' is a painter's pigment.
Not all color needs a band gap tuned to the visible; a second route is impurity levels. Pure alumina (sapphire's parent, Al2O3) is colorless, with a gap of about 9 eV. Drop in a pinch of chromium — a few chromium atoms replacing aluminium — and those impurity atoms plant new energy levels inside the wide gap. Electrons hop up to those levels by absorbing green and violet light, and what escapes is deep red: that impurity trick is a ruby, and swapping the chromium for iron and titanium gives the blue of sapphire. Even metals play this game: gold and copper look yellow and rosy rather than silver because their electrons absorb a little of the blue end before reflecting, tinting the mirror.
Luminescence: handing the light back
Absorption kicks an electron up to a higher energy level. What happens next? Often the energy just leaks away as heat, jiggling the atoms. But in some materials the electron falls back down across the gap and pays its energy back as a fresh photon of light. That re-emission is luminescence, and it is the engine behind glowing screens, fluorescent tubes, and glow-in-the-dark stars. There is a lovely twist: the electron usually tumbles down a couple of small steps as heat before the big light-emitting drop, so the photon it emits carries less energy — a longer wavelength — than the one it absorbed. That downhill shift is why a material can drink invisible ultraviolet and give back visible light.
This is exactly how a fluorescent tube lights your kitchen. Inside, an electrical discharge in mercury vapor pours out ultraviolet you cannot see; the white powder coating the glass — a phosphor — absorbs that UV and luminesces it back out as visible white light. The very same trick hides inside a 'white' LED: the chip itself emits only blue, and a yellow phosphor dabbed on top absorbs part of that blue and re-emits yellow, so blue-plus-yellow reaches your eye as white. Once you know to look, luminescence is everywhere the light seems to come from nowhere.
The LED: light straight out of a voltage
Luminescence started with a photon coming in. An LED skips that and starts with a voltage. Recall the p-n junction from the electrical rung: press it in forward bias and electrons pour in from the n-side while holes flood up from the p-side, meeting in the middle. When an electron drops into a hole it falls from the conduction band down to the valence band, and — in the right material — it lets that lost gap-energy go as a single photon. Millions of these tiny recombinations every instant add up to a steady glow. Because the light is born from an electric current, this is called electroluminescence, and the device is a light-emitting diode.
The color is set directly by the size of the band gap, through the same E = 1240 / wavelength we used for absorption — because the emitted photon carries almost exactly the gap energy. A gallium-arsenide junction with a 1.4 eV gap emits at 1240/1.4 ≈ 886 nm, invisible infrared, which is what the LED in your TV remote quietly flashes. Widen the gap toward 2.0 eV and you get red; push it near 2.7 eV and you get blue. That blue was the hard one: engineers had red and green LEDs for decades but could not make a bright blue until gallium nitride was tamed in the 1990s — and only then, red plus green plus blue, could LEDs make white light and light our world. It won a Nobel Prize.
Here is the honest catch that keeps LEDs off the world's favourite semiconductor. Silicon, the backbone of every chip, has an indirect band gap: an electron cannot simply fall across it and emit a photon, because the jump also demands a change in momentum that a photon alone cannot supply. So in silicon the recombination energy almost always leaks out as heat instead of light — silicon is a hopeless light-emitter. Efficient LEDs are therefore made from direct-gap compounds like gallium arsenide and gallium nitride, where the electron can drop straight down and hand its whole energy to a photon. It is a sharp reminder that a semiconductor's band gap is not just a number for its size; its shape decides whether the material can glow at all.
The laser: photons marching in step
An LED throws its photons out at random, each born whenever an electron happens to fall — a warm, scattered, many-colored glow. A laser does something radically more disciplined. Its trick is stimulated emission: when a photon passes an already-excited atom, it can coax that atom to emit a second photon that is an exact clone — same wavelength, same direction, and marching perfectly in phase, crest-on-crest, with the first. One photon becomes two identical ones, those become four, and an avalanche of perfectly synchronised light builds up. A common misconception is that a laser is just a very bright light; the real magic is that the light is coherent (all the waves in lockstep), monochromatic (one pure color), and collimated (a pencil-thin beam that barely spreads).
- Pump energy in. A flash lamp, an electric current, or another laser lifts a great many atoms up into an excited energy level — far more than would normally sit up there.
- Reach population inversion — the strange state where more atoms are excited than remain in the ground state. Without this over-full upper level, stimulated emission cannot outrun ordinary absorption.
- One atom drops on its own and emits a stray photon. As that photon sweeps past its excited neighbours, it triggers each to emit an identical clone — the chain reaction begins.
- Bounce it between two mirrors. The growing beam is trapped between a mirror at each end, sweeping back and forth through the excited medium and amplifying on every pass.
- Let a sliver escape. One mirror is made slightly leaky, and the thin thread of coherent light that streams out through it is your laser beam.
The first working laser, in 1960, was a rod of ruby — the very same chromium-doped alumina whose red we met a moment ago — flashed by a bright lamp. Today the workhorses are semiconductor laser diodes: an LED-like junction built inside a tiny mirrored cavity, so cheap and small that one sits in every barcode scanner and Blu-ray player. And crucially, a laser diode is the transmitter that fires crisp, monochromatic pulses of infrared into the last actor in our story — the optical fiber.
The optical fiber: light that carries the internet
How do you pipe a laser's pulses across an ocean without them dribbling out the sides? The answer leans entirely on total internal reflection, which you met in the last guide: when light travelling inside a high-index medium strikes the boundary with a lower-index one at a shallow enough angle, none of it escapes — it reflects with essentially 100 percent efficiency, a perfect mirror made of nothing but a step in the refractive index. An optical fiber is a hair-thin thread of ultra-pure glass with exactly that trick built in: a central core of slightly higher-index glass wrapped in a cladding of slightly lower-index glass.
OPTICAL FIBER: light trapped by total internal reflection
cladding (lower index, n ~ 1.46)
==========================================
/\ /\ /\ /\
light -> / \ / \ / \ / \ -> light out
/ \ / \ / \ / \
==========================================
cladding (lower index)
core (higher index): the ray meets the wall beyond the
critical angle, so ~100% reflects and zig-zags for km
instead of leaking out the side.
Loss ~ 0.2 dB/km at 1550 nm -> half the power every ~15 kmWhy glass, and why that particular infrared color? Because purity plus wavelength buys almost unbelievable transparency. Modern silica fiber loses only about 0.2 decibels per kilometre at a wavelength near 1550 nm — meaning the light still keeps half its power after roughly 15 km, and a useful fraction after 50 to 100 km, at which point an amplifier gives it a fresh push. Compare that to ordinary window glass, which would swallow your signal in a few metres. The whole game is beating absorption down: the 1970s breakthrough that made fiber possible was learning to purge the glass of iron and water down to a few parts per billion, because even a trace of those impurities plants absorbing levels that would drink the signal.