Saramsha

No. 034  ·  20 October 2026

The Angle Light Won't Cross

Light bends every time it crosses from one transparent medium into another — except when it's leaving a denser one at too steep an angle from the normal, and then it doesn't cross at all. Choose what the light is leaving and entering, drag the angle up, and watch the exit ray swing toward the surface and vanish.

exit angle, this pair no bending at all no exit — reflects instead

About this piece

Light doesn't slow down and speed up for fun

A beam of light crossing from one transparent material into another changes speed — slower in water than in air, slower still in glass, slower again in diamond — and that change of speed is the entire reason it bends. Picture a wide beam, really a wavefront, arriving at the boundary at an angle rather than straight on: one edge of it reaches the slower medium first and is held back while the rest is still travelling at the old speed, so the whole front pivots, the way a marching line bends when one end steps from a hard road onto soft mud while the other end is still marching at full pace. That pivot is refraction, and in 1621 Willebrord Snell pinned the exact amount of it to one equation: n₁ sin θ₁ = n₂ sin θ₂, where n is each material's refractive index — its speed penalty relative to a vacuum — and θ is the ray's angle from the normal, the line perpendicular to the boundary.

Run that equation one way and there's no drama. Light moving into a denser medium (lower speed, higher n) always bends toward the normal, and it can do that at any entry angle at all, right up to grazing the surface at 90°. But run it the other way — light trying to leave a denser medium for a thinner one, water into air, glass into air, diamond into air — and the exit angle has to grow faster than the entry angle, because now sin θ₂ = (n₁/n₂) sin θ₁ is being multiplied up rather than down. Raise θ₁ steeply enough and the exit angle reaches 90° — the refracted ray skims flat along the surface — while θ₁ itself is still comfortably short of 90°. That's the critical angle. Push past it and Snell's equation has no solution left to give: sin θ₂ would have to exceed 1, which no real angle produces. Nature's actual answer, confirmed in any swimming pool, is that the light simply doesn't leave. It reflects instead, off the inside of the surface, with almost no loss at all — cleaner than most mirrors manage, because nothing here is a coating absorbing a slice of the light; the boundary itself refuses to transmit any more.

Where you've actually seen this

An optical fibre is a glass thread carrying light down its length for kilometres with barely any loss, and the trick is entirely this cutoff: light is launched down the core at a shallow enough angle from the fibre's own axis that, every single time it reaches the boundary with the surrounding cladding, it is well past the critical angle. Every bounce is total, so none of it ever gets out through the sides — only distance and absorption in the glass itself cost it anything. A diver looking straight up sees the whole sky, horizon to horizon, squeezed into one bright cone directly overhead — roughly 97° wide, because the water-to-air critical angle is about 48.6° — and everywhere outside that cone the surface above them turns into a mirror, silently reflecting the pool floor or seabed back down, because any light trying to leave at a shallower angle than that is undergoing exactly the total internal reflection modelled here. And a diamond's unusually small critical angle, about 24.4°, is most of why a well-cut one looks so bright: its lower facets are angled so that light entering through the top is well past that narrow cutoff and has nowhere to go but back up, out through the top, toward the eye looking at it — a cheaply cut stone with the wrong facet angles lets light leak out the back instead, and looks duller for it.

Each material is modelled with one fixed refractive index (air 1.00, water 1.33, glass 1.50, diamond 2.42), ignoring the small amount real glass and diamond vary by colour — the actual source of a diamond's rainbow "fire", on top of the brightness effect shown here. The diagram also always sketches a faint partial reflection below the critical angle, as a reminder that a real boundary never transmits quite 100% of the light even when it isn't total; it just never draws the exact Fresnel fraction, which depends on polarisation too. ?from=, ?to= and ?angle= set the scene directly.