Lambert's Cosine Law — Why Incidence Angle Decides UV Dose

Quick Answer

When UV radiation strikes a surface at an angle, the irradiance actually delivered to that surface falls off with the cosine of the incidence angle. Hit a surface head-on (0°) and you get the full intensity; let the beam graze it (toward 90°) and the same energy spreads over an ever-larger area until the local intensity drops to zero.

This is Lambert's cosine law, first described by Johann Heinrich Lambert in his Photometria (1760). It is the reason industrial UV systems are positioned with care — and the reason a tilted lamp, a curved workpiece, or the edge of a coating can quietly receive far less dose than the centre.

What Lambert's Cosine Law States

When UV radiation strikes a surface obliquely, the effective irradiance on that surface scales with the cosine of the incidence angle:

I_surface = I_incident × cos(θ)

where θ is the angle between the ray direction and the surface normal (the perpendicular to the surface).

• At normal incidence (θ = 0°), cos(θ) = 1 — full intensity.
• At grazing incidence (θ → 90°), cos(θ) → 0 — the same flux is smeared across an effectively infinite projected area, and the local intensity collapses to zero.

The mechanism is purely geometric: a beam of fixed cross-section illuminates a larger patch of surface as the angle increases, so the energy per unit area drops by exactly the cosine factor. The same principle governs why a Lambertian (ideal diffuse) surface appears equally bright from every viewing direction — the radiant intensity it emits also follows the cosine of the angle to the normal.

Why This Matters in the UV Industry

This is not an academic footnote — it is a primary reason industrial UV systems are not always aimed straight down at the target.

Example 1: Bonding a Cylindrical Part

Consider a cylindrical part — a needle hub, a pin, a small shaft — with a UV adhesive bead, and a single point source directly above it. If the emitter shines straight down, it strikes the cylindrical sidewall almost tangentially: the incidence angle on the sides approaches 90°, the cosine factor there is small, and most of the UV passes by the part instead of being deposited where the adhesive sits.

A common solution is to illuminate the part from oblique angles with two or more emitters, so the sidewalls are hit closer to normal incidence. A documented variant for curing three-dimensional cylindrical objects without rotation uses multiple focused UV units oriented with their long dimension at a steep angle to the cylindrical axis, so the beam reaches surfaces that a single overhead lamp would miss.

Example 2: Conveyor Line Sources

A line source mounted over a moving web is the workhorse of UV curing. Aiming it perfectly vertically can send specularly reflected UV straight back into the lamp's own reflector, contributing to heat build-up in the housing.

Giving the source a slight tilt lets reflections clear the reflector while the substrate dose changes only marginally — because at small tilt angles the cosine factor is very close to 1 (see the table below). The small cosine loss is often an acceptable trade for reduced back-reflection into the lamp housing.

Example 3: Large-Area Coatings (Multi-Source Arrays)

When several UV bars are placed side by side to cover a wide substrate, the sources at the edge of the array illuminate the outer substrate edges at an angle. There the cosine factor is below 1, so the delivered dose drops — an edge falloff in cure. As a direct consequence of cos(θ) < 1 at the perimeter, the edge regions are where under-cure is most likely to appear.

A common mitigation is to keep the substrate narrower than the lamp array, so that the regions that matter sit in the zone where every source hits them close to normal incidence.

Practical Rule of Thumb: How Much Loss at Which Angle?

The cosine factor is exact trigonometry — it can be read straight off the cosine function:

Rule of thumb: up to about 20° of tilt the cosine loss stays under 10 % and is barely measurable. Beyond 30° the loss becomes significant (>13 %) and should be carried into the exposure-time or dose calculation. Beyond 45° the loss is severe and is only used deliberately — for example when the tilt itself is what lets the beam reach an awkward geometry such as a cylindrical sidewall.

A Common Trap: Inverse-Square vs. Cosine

The cosine law and the inverse-square law are two separate geometrical-optics effects, and they superpose:

I_surface = I_lamp × cos(θ) / r²
                     ↑           ↑
              Lambert cosine   inverse-square (distance²)

• Inverse-square law describes geometric spread: photons from a point source spread over a sphere whose area grows with r², so the local intensity falls as 1/r². Double the distance, quarter the irradiance.
• Lambert's cosine law describes hit efficiency: obliquely incident photons spread their energy over a larger patch of surface, so the local intensity falls as cos(θ).

Accounting for both at the same time is standard practice in any serious irradiance-distribution calculation — and not only in UV. The identical principle appears in photometry, computer-graphics rendering, antenna engineering, and solar-collector design.

Designing Around the Cosine Law

When laying out a UV installation, two practical takeaways follow directly from the law:

1. Specify the worst-case point. The location of the workpiece that is hit at the largest incidence angle (or the greatest distance, or in shadow) sets the dose for the whole job. Curing or disinfection at that point must still meet the target.
2. Several oblique sources can beat one perpendicular source. For curved or three-dimensional workpieces, two or three sources arriving from different directions often deliver a more uniform surface dose than a single overhead lamp — each source covers the facets the others hit at a poor angle.

A useful customer-facing note for tilted installations:

When a UV lamp is tilted relative to the target, the surface irradiance drops according to Lambert's cosine law. Up to roughly 20° of tilt the effect is negligible (under 10 % loss). Beyond about 30°, account for it in the exposure-time calculation or specify a correspondingly higher lamp power. For cylindrical or curved workpieces, two or more lamps at different angles usually give a more even surface dose than a single perpendicular lamp.

Cross-References

• Layer Thickness & Dose Scaling — how delivered surface dose translates into cure through a coating
• Reflector Geometries — how reflector shape redirects UV and interacts with incidence angle
• UV System Type Taxonomy — where surface-dose vs. volume-dose modelling applies
• Wavelengths & Action Spectra — the spectral side of effective dose

Sources

• Lambert's cosine law — Wikipedia (definition, Lambertian reflectance, historical origin in Lambert's Photometria, 1760)
• "Cosine law" — ScienceDirect Topics (irradiance vs. incidence angle in radiometry)
• Lambert's Cosine Law — optris knowledge library (angular dependence of emitted/incident radiation)
• "Inverse Square Law of Irradiance" — SPIE Field Guide to Radiometry (point-source 1/r² behaviour)
• Light Measurement Handbook, Basic Principles — International Light Technologies (inverse-square and cosine law as the two laws of geometrical optics for radiometry/photometry)
• US Patent 4,208,587 — "Method and apparatus for ultraviolet curing of three-dimensional objects without rotation" (multiple focused UV units angled to a cylindrical axis)