Polarization Lab · Guide · glidepathscience.org
GuidePolarization Lab
Contents
This Guide covers the ideas and equations behind every Walkthrough step, Concept Challenge question and Predict round. Read Part 1 before the Intro track and Part 2 before the Advanced track, about ten minutes each. Angles are measured from the horizontal, counterclockwise as seen looking into the beam, so 90° is vertical; the polarizers are ideal unless a section says otherwise.
Part 1 · The essentials
Light, and the ½ rule
Light is a transverse wave: its electric field $\vec{E}$ oscillates across the direction the light travels. The direction it oscillates in is the light’s polarization. In linear light $\vec{E}$ swings along one line; in circular and elliptical light it turns. Light from the Sun, a flame or a bulb is unpolarized: its field changes direction randomly and far too fast to follow. Only transverse waves can be polarized; sound in air cannot.
An ideal polarizer passes the part of $\vec{E}$ along its transmission axis and removes the part across it. Unpolarized light has no special direction, and averaging $\cos^2\theta$ over every direction gives ½, so whatever the polarizer’s angle
$I = \tfrac12 I_0$
The light that leaves is polarized along the axis. Turning the polarizer turns the light that leaves, not how much of it there is.
What a polarizer passes
A wire-grid polarizer is a row of fine parallel metal wires, closer together than a wavelength. The field along the wires drives electrons up and down them, and that part of the light is absorbed or reflected. The field across the wires cannot drive a current across a thin wire, so it passes. The transmission axis is therefore perpendicular to the wires: the gaps are not slots that let light “slip through” along them. Polarizer film works the same way, with long aligned molecules for wires. In this sim every axis is drawn as a double arrow, the direction of $\vec{E}$ that passes.
Malus’s law
Polarized light meets an analyzer whose axis is at θ to its polarization. Only the component along the axis, $E\cos\theta$, gets through, and a meter reads intensity, which goes as the field squared:
$I = I_0\cos^2\theta$
θ is the angle between the light’s polarization and the axis, not the axis’s angle from the horizontal. There is no ½: the light is already polarized. And the cosine is squared, so the fall is flat near 0°, steepest at 45° and zero at 90°, not a straight line.
Chains, and crossed polarizers
A chain of polarizers on a lamp takes two rules, in order: the ½ at the first, because the lamp’s light is unpolarized; then Malus’s law at each one after it, with the angle between neighboring axes:
$I_k = I_{k-1}\cos^2(\theta_k-\theta_{k-1})$
Polarizers 90° apart are crossed: the light leaving the first has no component along the second, and nothing gets through. Order matters: 0°, 90°, 45° passes nothing, while 0°, 45°, 90° does pass light.
A third polarizer, and many gentle steps
Slip a polarizer between a crossed pair, at θ from the first. The light now turns in two steps, θ and 90° − θ, and some of it gets through:
$I = \tfrac18 I_0\sin^2 2\theta$
The most, $I_0/8$, comes at 45°. A polarizer is not a sieve that only removes light: it re-projects the field onto its axis, and the last polarizer then receives light it can pass. Between any two axes φ apart, the middle one does best halfway, at φ/2: equal steps lose the least. N polarizers across 90° in equal steps pass
$\tfrac12\cos^{2(N-1)}(90^\circ/(N-1))$
of a lamp’s light, approaching ½ as N grows: turning a polarization gently loses little.
Glare and Brewster’s angle
Reflection polarizes light. Take the plane that holds the incoming, reflected and refracted rays: the plane of incidence, drawn as the page. The s part of the field is perpendicular to it, parallel to the surface (drawn ⊙ and ⊗); the p part lies in it (drawn as arrows). For light reflecting off water or glass, s reflects more than p at every angle between 0° and 90°, so glare from a lake or a road is mostly horizontal, and polarized sunglasses have vertical axes. At one angle p does not reflect at all, Brewster’s angle:
$\tan\theta_B = n_2/n_1$
There the reflected and refracted rays are 90° apart, and unpolarized light reflects as pure s.
The polarized sky
Air molecules, far smaller than a wavelength, scatter sunlight as tiny dipoles driven by its field, and a dipole radiates nothing along its own axis. Looking 90° from the Sun you see only the part of their motion across your line of sight, so that light is strongly polarized; toward and away from the Sun it is barely polarized. With the Sun low in the west, the most polarized band runs overhead, through north and south. The same dipoles scatter short wavelengths more, as $\lambda^{-4}$: blue light at 450 nm about 4.4 times as much as red at 650 nm.
Part 2 · Going further
Jones vectors and matrices
Fully polarized light is a Jones vector, $\vec{E} = (E_x, E_y)$, with complex components that carry the phase. Each element is a 2 × 2 matrix; a polarizer with its axis at θ is
$P(\theta) = \begin{pmatrix} \cos^2\theta & \sin\theta\cos\theta \\ \sin\theta\cos\theta & \sin^2\theta \end{pmatrix}$
A stack is one matrix, $\vec{E}_{out} = P_n\cdots P_2P_1\vec{E}_{in}$, with the first element met on the right, and $I = |E_x|^2+|E_y|^2$. Matrices do not commute, which is the algebra behind “order matters”.
Real polarizers, and measuring them
A real sheet passes a fraction $k_1$ of light polarized along its axis and leaks $k_2$ across it. On a lamp, one sheet passes $(k_1+k_2)/2$, a parallel pair $\tfrac12(k_1^2+k_2^2)$ and a crossed pair $k_1k_2$; at any angle,
$I/I_0 = k_1k_2 + \tfrac12(k_1-k_2)^2\cos^2\theta$
Malus’s law on a floor. On the Lab Bench, fit $A\cos^2(\theta-\theta_0)+B$ to a sweep: $\theta_0$ is the mount reading where the axes line up, and the floor B also holds the meter’s dark reading D, so the extinction, floor over peak, is $(B-D)/(A+B-D)$; for two identical sheets it is $2k_1k_2/(k_1^2+k_2^2)$.
Fresnel’s equations, and why p vanishes
With Snell’s law, $n_1\sin\theta_i = n_2\sin\theta_t$, the reflected amplitudes are (with Hecht’s signs; some texts flip $r_p$)
$r_s = (n_1\cos\theta_i-n_2\cos\theta_t)/(n_1\cos\theta_i+n_2\cos\theta_t)$ $r_p = (n_2\cos\theta_i-n_1\cos\theta_t)/(n_2\cos\theta_i+n_1\cos\theta_t)$
and the reflected intensities $R = r^2$. For external reflection $R_s \ge R_p$, equal at normal incidence, and $R_p$ = 0 at $\theta_B$. The reason is the dipoles: the reflected light is radiated by electrons in the glass, driven by the refracted wave, and for p light they oscillate in the page, across the refracted ray. When the two rays are perpendicular, $\theta_B + \theta_t = 90°$, the reflected direction lies along their motion, where a dipole sends nothing. Light inside glass has its own Brewster angle, with $n_1$ the glass, below the critical angle beyond which it all reflects.
A pile of plates
A stack of glass plates at Brewster’s angle polarizes by transmission: p passes every surface, while s loses $R_s$ at each of the 2m surfaces of m plates. Some of the reflected s light bounces back and forth and comes out after all; counting those reflections (incoherently, as for thick plates), each polarization passes
$T = (1-R)/(1+(2m-1)R)$
and the transmitted light’s degree of polarization is $(T_p-T_s)/(T_p+T_s)$, the same as Hecht’s $m/(m+(2n/(1-n^2))^2)$.
Wave plates
A birefringent plate has two refractive indices, so the component along its slow axis falls behind by the retardance $\Gamma = 2\pi\,\Delta n\,d/\lambda$. A half-wave plate ($\Gamma = \pi$) mirrors linear light in its fast axis φ, without loss: $\alpha \to 2\varphi-\alpha$. A quarter-wave plate ($\Gamma = \pi/2$) with its fast axis horizontal takes $(\cos\alpha, \sin\alpha) \to (\cos\alpha, i\sin\alpha)$: a quarter wave apart, so light at 45° comes out circular, and an analyzer at β then passes $\cos^2\alpha\cos^2\beta + \sin^2\alpha\sin^2\beta$. Right-circular light turns clockwise looking into the beam (Hecht’s naming). Through an analyzer alone, circular light reads a steady half, as unpolarized light does; a quarter-wave plate first turns it linear, and then the analyzer can extinguish it.
Stokes parameters
Partly polarized light has no Jones vector, but four intensities a meter can measure describe it: $S_0 = I_H+I_V$, $S_1 = I_H-I_V$, $S_2 = I_{45}-I_{135}$ and $S_3 = I_R-I_L$, where $I_R$ and $I_L$ are read through a quarter-wave plate (fast axis horizontal) and an analyzer at 45° and 135°. The degree of polarization is $\sqrt{S_1^2+S_2^2+S_3^2}/S_0$, and the azimuth $\tfrac12\operatorname{atan2}(S_2, S_1)$, which has no meaning for circular or unpolarized light. No passive element raises $S_0$, and always $S_0^2 \ge S_1^2+S_2^2+S_3^2$.
Rayleigh scattering, measured
For an ideal dipole in unpolarized light, the part of the field perpendicular to the scattering plane scatters fully and the part in it as $\cos^2\psi$, so at scattering angle ψ
$P(\psi) = \sin^2\psi/(1+\cos^2\psi)$
which is 1 at 90°. Air’s molecules are slightly anisotropic, with a depolarization ratio ρ ≈ 0.028, which caps P near 0.95; multiple scattering, aerosols and light from the ground bring a real clear sky’s maximum to roughly 0.7 to 0.8.
A screen pixel
In a twisted-nematic pixel, liquid-crystal molecules twist through 90° from one face of the cell to the other, and the light’s polarization follows the twist: the continuous limit of many gentle steps, without their loss. Between crossed polarizers the pixel is therefore bright with no voltage. A voltage stands the molecules up along the field, nothing twists, and the pixel goes dark. A screen’s light leaves through its front polarizer, which is why it dims through polarized sunglasses turned across it.
Reference
Common mistakes
- Malus’s law on a lamp. Unpolarized light has no direction to measure θ from: ½, at any angle.
- Losing the ½, or adding one for light that is already polarized.
- cos instead of cos². Backward, take the square root first: θ = arccos √f, not arccos f.
- The wrong angle. Use the angle between neighboring axes, not each axis’s angle from the horizontal.
- “A polarizer is a sieve.” It re-projects the field, so a third one brings light back.
- Slots. A wire grid passes the field across its wires.
- Brewster by arcsine, by the ratio upside down, or in radians.
- Amplitudes for intensities. $R = r^2$, and the meter reads intensity.
- Circular is not unpolarized. Both read ½ through an analyzer; a quarter-wave plate tells them apart.
Equation sheet
| Part 1 | |
|---|---|
| lamp → polarizer | $I = \tfrac12 I_0$, any angle |
| Malus’s law | $I = I_0\cos^2\theta$ |
| chain | $I_k = I_{k-1}\cos^2(\theta_k-\theta_{k-1})$ |
| between crossed | $I = \tfrac18 I_0\sin^2 2\theta$ |
| N across 90° | $\tfrac12\cos^{2(N-1)}(90^\circ/(N-1))$ |
| Brewster | $\tan\theta_B = n_2/n_1$ |
| Part 2 | |
|---|---|
| real pair | $I/I_0 = k_1k_2 + \tfrac12(k_1-k_2)^2\cos^2\theta$ |
| Snell | $n_1\sin\theta_i = n_2\sin\theta_t$ |
| Fresnel | $r_s = (n_1\cos\theta_i-n_2\cos\theta_t)/(n_1\cos\theta_i+n_2\cos\theta_t)$, $R = r^2$ |
| $r_p = (n_2\cos\theta_i-n_1\cos\theta_t)/(n_2\cos\theta_i+n_1\cos\theta_t)$ | |
| pile of m plates | $T = (1-R)/(1+(2m-1)R)$; degree $(T_p-T_s)/(T_p+T_s)$ |
| half-wave plate | $\alpha \to 2\varphi-\alpha$ |
| quarter-wave plate | $(\cos\alpha, \sin\alpha) \to (\cos\alpha, i\sin\alpha)$ |
| retardance | $\Gamma = 2\pi\,\Delta n\,d/\lambda$ |
| Stokes | $\sqrt{S_1^2+S_2^2+S_3^2}/S_0$ |
| Rayleigh | $P(\psi) = \sin^2\psi/(1+\cos^2\psi)$ |
Symbols: I intensity (W/m²), $I_0$ at the first element · θ angle between a polarization and an axis · $k_1$, $k_2$ a sheet’s transmittances along and across its axis · n refractive index, $\theta_i$, $\theta_t$ incidence and refraction · s, p perpendicular to and in the plane of incidence · $r$ amplitude, R intensity reflected, T transmitted · m number of plates · φ a plate’s fast axis, α the light’s angle · Γ retardance, Δn birefringence, d thickness · $S_0$…$S_3$ Stokes parameters · ψ scattering angle, ρ depolarization ratio.
Further reading
E. Hecht, Optics, 5th ed. (Pearson, 2017): Malus’s law, Fresnel’s equations, Brewster’s angle, the pile of plates, retarders, and Jones and Stokes calculus. OpenStax University Physics Volume 3, §1.7 (Polarization), and College Physics 2e, §27.8, free online. And this sim’s Field Manual (H): how the model works, and what it leaves out. For one photon at a time, see the Quantum Eraser Lab.
Polarization Lab v1.2 · © 2026 Dr. David Meier · CC BY-NC-SA 4.0 · glidepathscience.org