Light hits your retina right now. Photons that left the sun eight minutes ago just passed through your cornea, your lens, the vitreous humor, and slammed into photoreceptors that turn physics into the words you're reading Small thing, real impact..
Wild, right?
We don't think about transmission. We think about seeing. Which means glass. But everything visual — every color, every shadow, every pixel on this screen — starts with light waves moving through something. Water. Also, air. The clear tissue at the front of your eye That alone is useful..
Most people know light travels. Fewer know how it gets from there to here without turning into noise.
Let's fix that.
What Is Light Transmission
Transmission is just the passage of electromagnetic waves through a medium. In real terms, that's the textbook version. Here's the real one: light enters a material, interacts with its atoms, and — if the conditions are right — keeps going Still holds up..
Not all light makes it. Some reflects. Some absorbs. Some scatters. What survives is transmitted.
The medium matters. A lot. Vacuum? In real terms, zero resistance. On top of that, light hits 299,792,458 meters per second and doesn't slow down. Air? Close enough — 0.03% slower. Water? In practice, 25% slower. Glass? 33% slower. Diamond? Nearly 60% slower Practical, not theoretical..
That slowdown? Photons always move at c. Practically speaking, it's not the photons dragging. It's the absorption-reemission dance with electrons in the material. So each interaction adds a tiny delay. Billions of delays later, the wave emerges later than it would have in vacuum.
We call that ratio the refractive index. n = c / v. Simple number. Massive consequences.
The Two Flavors of Transmission
Transparent transmission — what you get from clean glass, pure water, clear air. Light passes with minimal scattering. You see through the medium.
Translucent transmission — frosted glass, thin paper, skin. Light gets through but scatters enough to blur detail. You see light, not images It's one of those things that adds up..
Opaque materials? Also, they transmit almost nothing in the visible range. But shine infrared through silicon — suddenly it's transparent. "Transparent" depends entirely on wavelength.
Why It Matters
You're using transmission right now. Fiber optic cables carry this article to your device as pulses of infrared light through ultra-pure silica glass. Total internal reflection keeps the signal trapped for kilometers. Without understanding transmission — absorption windows, dispersion, scattering losses — the internet doesn't exist.
Your eyes? Dual transmission systems. The cornea and lens transmit visible light while blocking UV. The retina's photoreceptors sit behind layers of neurons and blood vessels — light transmits through living tissue before hitting the sensors. Evolution solved the wiring problem by making the tissue transparent enough.
Cameras. Microscopes. Solar panels. Laser surgery. Here's the thing — telescopes. Every optical technology lives or dies by transmission physics.
And here's what most people miss: transmission isn't binary. A material might transmit 95% of green light but only 10% of blue. It's spectral. That's why cheap sunglasses distort colors — they don't transmit evenly across the spectrum It's one of those things that adds up..
How It Works
Light hits a boundary. Three things happen simultaneously That's the part that actually makes a difference..
Reflection at the Interface
Every boundary reflects some light. Air to glass? Because of that, ~4% per surface at normal incidence. That's why uncoated lenses lose 8% right off the bat — 4% front surface, 4% back surface Simple, but easy to overlook..
The Fresnel equations govern this. But they depend on refractive indices, angle of incidence, and polarization. At Brewster's angle, p-polarized light transmits 100% — zero reflection. That's how polarizing filters work.
Absorption Inside the Medium
Photons meet electrons. If the photon energy matches an electronic transition — boom, absorbed. The electron jumps. The photon vanishes The details matter here. Turns out it matters..
In glass, visible photons don't match UV electronic transitions or IR vibrational modes. So they pass through. But add iron impurities? The iron absorbs red and blue. Green tint. That's why old window glass looks green at the edges — thicker path, more absorption.
Absorption follows Beer-Lambert law: I = I₀e^(-αx). 2 dB/km. Double the thickness, square the loss. Exponential decay. This is why submarine fiber uses absurdly pure glass — attenuation below 0.Regular window glass would kill the signal in meters.
Scattering: The Silent Thief
Rayleigh scattering — particles much smaller than wavelength. But scales as 1/λ⁴. Blue scatters 16x more than red. That's why the sky is blue and sunsets are red. Same physics makes optical fibers lose more at shorter wavelengths Worth keeping that in mind..
Mie scattering — particles comparable to wavelength. Wavelength-independent. Fog, clouds, milk. This kills transmission brutally.
In high-quality optics, scattering comes from surface roughness and bulk inhomogeneities. Polish matters. Purity matters Small thing, real impact..
Refraction and the Speed Change
Light enters at an angle. Even so, one side slows first. The wavefront hits the boundary at different times. The wave bends.
Snell's law: n₁sinθ₁ = n₂sinθ₂. Simple. Universal.
But n isn't constant. That's why prisms split white light. It varies with wavelength — dispersion. Blue bends more than red. And lenses focus different colors at different points — chromatic aberration. Camera lenses fight this with achromatic doublets: two glasses with opposing dispersion.
Group velocity vs phase velocity matters for pulses. Data rates hit limits. That's why pulses spread. Even so, in fiber, different wavelengths travel at different group velocities — group velocity dispersion. Dispersion compensation modules fix it.
Common Mistakes
"Light slows down in glass."
Photons don't slow. The wave appears slower because of absorption-reemission delays. The distinction matters for Cherenkov radiation — particles can exceed light's phase velocity in a medium, creating the blue glow in nuclear reactors.
"Transparent means 100% transmission."
No. Even the best anti-reflection coated glass transmits ~99.5% per surface. Two surfaces = 99%. Ten lens elements = ~90% throughput. Photographers know this. T-stops exist because f-stops lie about transmission Turns out it matters..
"Thicker glass just means more absorption."
Sometimes. But scattering and reflection losses also scale with thickness. And thermal lensing — high-power lasers heat the glass, changing n, creating a lens inside the material. This destroys beam quality in high-energy systems That alone is useful..
"Air is basically vacuum for optics."
At 1 atm, n = 1.000277. Over 1 km, that's 277 microns of optical path difference. Interferometers care. Adaptive optics care. LIGO cares — they operate in ultra-high vacuum for a reason.
"Polarization doesn't affect transmission."
At non-normal incidence, s and p polarizations reflect differently. At Brewster's angle, p-polarization transmits perfectly. Laser cavities use Brewster windows for this exact reason — zero loss for the desired polarization.
Practical Tips
Coat everything.
Single-layer MgF₂ cuts reflection from 4% to ~1.5%. Multi-layer broadband AR coatings hit <0.25% across the visible. If you're building an optical system and skipping coatings, you
are leaving 10–15% of your signal on the table. Broadband AR is cheap insurance. In real terms, narrowband V-coats hit <0. 1% at design wavelength — essential for laser lines Took long enough..
Index-match when you can.
Oil immersion objectives (n=1.515) eliminate the air-glass interface. Transmission jumps. Resolution jumps. Same principle: couple fiber to detector with index-matching gel. Fresnel loss vanishes.
Watch your angles.
Normal incidence is forgiving. At 45°, unpolarized reflection jumps to ~10% per surface. P-polarization at Brewster's angle? Near zero. Design for it. Polarizing beamsplitters require it Most people skip this — try not to..
Clean like your data depends on it — it does.
Fingerprints scatter. Hydrocarbons absorb IR. Particulates nucleate laser damage. Use lint-free wipes, reagent-grade solvent, single-direction strokes. Re-inspect under dark-field illumination. If you see streaks, do it again.
Thermal management is optical management.
High average power? Water-cooled mounts. Low-expansion substrates (Zerodur, ULE). Active thermal control to ±0.01°C. A 1°C gradient in a 100mm optic induces wavefront error exceeding λ/4. That's diffraction-limited performance gone.
Specify transmission at your wavelength, not the catalog peak.
"High transmission 400–700nm" means nothing at 1064nm or 1550nm. Get the spectral curve. Measure it if the coating run matters. Batch variation is real Surprisingly effective..
Don't forget the cement.
Doublets use epoxy or frit. Both absorb. Both yellow under UV. Both have refractive index mismatches at the bond line. For high power or deep UV: air-spaced, optically contacted, or diffusion-bonded. Cement is a compromise — know when you've accepted it Practical, not theoretical..
The Bottom Line
Transmission isn't a single number. It's a spectral curve, an angular function, a polarization dependence, a thermal history, and a surface-quality map all at once Worth knowing..
Every interface steals. Every impurity scatters. Every gradient distorts.
The systems that win — the lithography steppers printing 3nm features, the gravitational wave detectors measuring 10⁻¹⁸m displacements, the free-space optical links pushing terabits through atmosphere — don't just "use good glass.And they coat every surface. And they control every temperature. So naturally, " They model every loss mechanism. They clean like surgeons.
Because in optics, what you don't lose is what you detect. And detection is the whole game.
