ARCHETYPE: Prism, Deviation & Dispersion Numericals

Newton's prism did two things at once, and every exam prism problem is one of them. It bent a beam of light toward its thick base — that is deviation — and it spread white light into a fan of colours, because it bends violet more than red — that is dispersion. Master the small handful of relations that quantify those two effects and an entire family of CBSE, JEE and NEET questions collapses into fill-in-the-blanks: find the refractive index from the minimum deviation, find the deviation of a thin prism, find the dispersive power of a glass, or design a two-prism combination that bends without colouring, or colours without bending. The keystone is the minimum-deviation relation μ=sin⁡ ⁣(A+δm2)/sin⁡ ⁣(A2)\mu = \sin\!\big(\tfrac{A + \delta_m}{2}\big)/\sin\!\big(\tfrac{A}{2}\big), which turns a protractor measurement of a prism's least deviation δm\delta_m into its refractive index μ\mu — a 60∘60^\circ crown prism that deviates a beam by at least 37.2∘37.2^\circ betrays a μ\mu of 1.51.5. For thin prisms the whole thing linearises to δ=(μ−1)A\delta = (\mu - 1)A, and from there the angular dispersion (μv−μr)A(\mu_v - \mu_r)A and the dispersive power ω=(μv−μr)/(μy−1)\omega = (\mu_v - \mu_r)/(\mu_y - 1) — a pure property of the glass — fall straight out. The combination problems are just two of these balanced against each other. This is a numericals gym, so the setup is the article. We fix one master recipe — name the quantity asked, pick the matching relation, and for combinations set the crown-against-flint balance — and then run it on the four canonical problem types plus the trap that catches the unwary: a prism apex too large transmits nothing at all, because the ray hits the second face beyond the critical angle. By the end, "A=60∘A = 60^\circ, δm=40∘\delta_m = 40^\circ" should make you reach for one formula without hesitation.

Key takeaways
  • Every prism numerical is one of four: (a) refractive index from minimum deviation, (b) thin-prism deviation, (c) dispersive power, or (d) a two-prism combination.
  • Minimum deviation: μ=sin⁡ ⁣(A+δm2)/sin⁡ ⁣(A2)\mu = \sin\!\big(\tfrac{A+\delta_m}{2}\big)/\sin\!\big(\tfrac{A}{2}\big). With A=60∘A = 60^\circ, δm=40∘\delta_m = 40^\circ: μ=sin⁡50∘/sin⁡30∘=1.532\mu = \sin 50^\circ/\sin 30^\circ = 1.532. With A=60∘A = 60^\circ, μ=1.5\mu = 1.5: δm=37.2∘\delta_m = 37.2^\circ.
  • Thin prism: δ=(μ−1)A\delta = (\mu - 1)A. A 5∘5^\circ crown prism (μ=1.5\mu = 1.5) deviates by 2.5∘2.5^\circ.
  • Dispersive power ω=(μv−μr)/(μy−1)\omega = (\mu_v - \mu_r)/(\mu_y - 1), a property of the glass alone. Crown (μr=1.514\mu_r = 1.514, μy=1.517\mu_y = 1.517, μv=1.523\mu_v = 1.523) gives ω=0.0174\omega = 0.0174.
  • Combinations: direct-vision (no net bend) balances the deviations (μy−1)A=(μy′−1)A′(\mu_y - 1)A = (\mu'_y - 1)A'; achromatic (no net colour) balances the dispersions (μv−μr)A=(μv′−μr′)A′(\mu_v - \mu_r)A = (\mu'_v - \mu'_r)A'.
  • Key idea: pick the relation from what is asked — minimum-deviation index, thin-prism deviation, dispersive power, or a crown-flint balance — and always check the prism apex is below the twice-critical-angle limit so light gets through at all.
The core insight

A prism both deviates a ray toward its base and disperses white light because μ\mu depends on colour; every numerical is one of four relations — μ=sin⁡(A+δm2)/sin⁡(A2)\mu = \sin(\tfrac{A+\delta_m}{2})/\sin(\tfrac{A}{2}) at minimum deviation, δ=(μ−1)A\delta = (\mu - 1)A for a thin prism, ω=(μv−μr)/(μy−1)\omega = (\mu_v - \mu_r)/(\mu_y - 1) for dispersive power, and the crown-flint balances for combinations — applied after checking the apex is below 2C2C so the ray escapes the second face.

Where this lives today

The minimum-deviation formula is not a museum piece — it is how the refractive index of an optical glass is measured, to five decimal places, on a spectrometer, by reading the least deviation of a prism cut from it. Every catalogue value of μ\mu began as an AA and a δm\delta_m. The dispersive-power balance is the working principle of the achromatic lens in every camera, microscope and pair of binoculars: a crown element and a flint element cemented together so their colour spreads cancel while their focusing powers do not, exactly equation (6) applied to lenses instead of prisms — the invention that made sharp, colour-free optics possible, which Chester Moore Hall first devised around 1730 and John Dollond patented in 1758, its priority contested in the courts in the 1760s. The direct-vision prism lives in pocket spectroscopes and hand-held mineral identifiers, letting you see an emission spectrum straight down the barrel. And the same dispersion, spread across a diffraction grating's successor, is how a spectrometer splits starlight or a chemical flame into the lines that reveal what it is made of — the science Fraunhofer's solar lines began.

The setup — one prism, two effects, four relations

A prism is a wedge of glass with a refracting angle (apex angle) AA. A ray entering one face is refracted toward the base, crosses to the second face, and is refracted again on the way out, emerging turned through a total angle of deviation δ\delta from its original path. Track the angles: the ray refracts from incidence ii to an internal angle r1r_1 at the first face, travels to the second face where it makes an internal angle r2r_2 and emerges at angle ee. Two facts of geometry govern everything. The internal angles fill the apex, r1+r2=Ar_1 + r_2 = A, and the total deviation is the sum of the two individual turns, δ=(i−r1)+(e−r2)=i+e−A\delta = (i - r_1) + (e - r_2) = i + e - A. As you rotate the prism, δ\delta first falls, reaches a minimum, then rises again — and the minimum occurs exactly when the passage is symmetric, i=ei = e and r1=r2=A/2r_1 = r_2 = A/2. That symmetry is what makes the minimum-deviation measurement so clean: it needs no knowledge of the internal path, only the apex angle and the least deviation you can read off a spectrometer. At that special geometry, δm=2i−A\delta_m = 2i - A, and Snell's law at the first face (quoted from earlier in this chapter) gives the master relation the whole topic hangs on.

Figure
A prism at minimum deviation: the ray passes symmetrically with equal angles at the two faces, deviated by the least amount, and because the index depends on colour the white light fans into a spectrum from red to violet
Prism at minimum deviation — symmetric passage, white light fans to a spectrum A refracting angle A white light, incidence i r₁ = r₂ = A/2 red (bent least) violet (bent most) original direction deviation δ_m μ = sin((A + δ_m)/2) / sin(A/2)

Here is the master recipe for every problem below. First, name the quantity asked — an index, a deviation, a dispersive power, or a combination angle. Second, pick the matching relation from the four. Third, for a combination, set the crown-against-flint balance — deviations equal for direct-vision, dispersions equal for achromatic. Fourth, check the apex limit A≤2CA \le 2C, because a prism whose apex is too wide throws the internal ray past the critical angle at the second face, where it total-internally-reflects instead of emerging, and no answer exists.

The four working relations

Everything numerical about a prism is in these four lines. At minimum deviation,

μ=sin⁡ ⁣(A+δm2)sin⁡ ⁣(A2)(1)\mu = \frac{\sin\!\left(\dfrac{A + \delta_m}{2}\right)}{\sin\!\left(\dfrac{A}{2}\right)} \qquad (1)

where: AA is the refracting angle, δm\delta_m the minimum deviation and μ\mu the refractive index. For a thin prism (small AA, so all angles small and sines equal their arguments), this collapses to the linear

δ=(μ−1)A(2)\delta = (\mu - 1)A \qquad (2)

independent of the angle of incidence — the deviation is just the "extra bending power" (μ−1)(\mu - 1) times the apex. Because μ\mu is larger for violet than for red, each colour gets its own deviation, and the spread between the extremes is the angular dispersion

θ=δv−δr=(μv−μr)A(3)\theta = \delta_v - \delta_r = (\mu_v - \mu_r)A \qquad (3)

while the colour-spread measured per unit mean deviation is the dispersive power

ω=μv−μrμy−1(4)\omega = \frac{\mu_v - \mu_r}{\mu_y - 1} \qquad (4)

where: μv\mu_v, μr\mu_r, μy\mu_y are the indices for violet, red and mean (yellow) light. Note what (4) does not contain: the apex AA cancels, so dispersive power is a pure property of the glass, dimensionless, the same for a fat prism or a thin one. Crown glass runs around ω=0.017\omega = 0.017, flint glass roughly double — the difference that makes achromatic optics possible. These four relations, plus the two combination balances, answer every prism numerical there is.

Combinations — bend without colour, or colour without bend

Put two thin prisms base-to-base of opposite orientation — a crown prism of angle AA and a flint prism of angle A′A' — and their deviations and dispersions each subtract. Two designs matter. A direct-vision prism spreads colour but leaves the mean beam undeviated, which needs the mean deviations to cancel:

(μy−1)A=(μy′−1)A′(5)(\mu_y - 1)A = (\mu'_y - 1)A' \qquad (5)

An achromatic combination bends the beam but leaves no residual colour, which needs the dispersions to cancel:

(μv−μr)A=(μv′−μr′)A′(6)(\mu_v - \mu_r)A = (\mu'_v - \mu'_r)A' \qquad (6)

Equation (5) is the spectroscope's direct-vision prism (see the spectrum straight ahead); equation (6) is the achromatic lens's secret (focus all colours together). Every combination problem is one of these two balances solved for the unknown flint angle A′A'.

Figure
Two combination designs from a crown and a flint prism: direct-vision cancels the net bend but keeps the colour spread, achromatic cancels the net colour but keeps the bend
Crown plus flint — two opposite designs DIRECT-VISION — no net bend, colour spreads crown A flint A' mean beam straight, colours fan (μ_y − 1)A = (μ'_y − 1)A' ACHROMATIC — no net colour, beam bends crown A flint A' one white beam, bent (μ_v − μ_r)A = (μ'_v − μ'_r)A' direct-vision balances the DEVIATIONS · achromatic balances the DISPERSIONS crown A = 6° with flint gives A' = 4.77° (direct-vision) or 2.7° (achromatic)
Derivation AThe minimum-deviation master solve, both directions

The goal is the relation μ=sin⁡ ⁣(A+δm2)/sin⁡ ⁣(A2)\mu = \sin\!\big(\tfrac{A+\delta_m}{2}\big)/\sin\!\big(\tfrac{A}{2}\big) and its two uses: getting μ\mu from a measured (A,δm)(A, \delta_m) — landing μ=1.532\mu = 1.532 for A=60∘A = 60^\circ, δm=40∘\delta_m = 40^\circ — and inverting it to get δm\delta_m from a known μ\mu — landing δm=37.2∘\delta_m = 37.2^\circ for A=60∘A = 60^\circ, μ=1.5\mu = 1.5. The strategy is pure geometry plus one use of Snell's law: track the ray through both faces, impose the symmetry that defines minimum deviation, and read off the index.

Step 1. Set up the angles. A ray enters the first face at incidence ii, refracts to internal angle r1r_1, crosses to the second face at internal angle r2r_2, and emerges at ee. Geometry of the triangle gives r1+r2=Ar_1 + r_2 = A, and the total deviation is the sum of the turns at the two faces, δ=(i−r1)+(e−r2)=i+e−A\delta = (i - r_1) + (e - r_2) = i + e - A.

Step 2. Impose the minimum. As the prism rotates, δ\delta passes through a least value, and at that point the path is symmetric: i=ei = e and r1=r2=A/2r_1 = r_2 = A/2. (Symmetry is forced because the deviation curve is smooth and the two faces are interchangeable.) Then δm=2i−A\delta_m = 2i - A, so the angle of incidence at minimum deviation is i=A+δm2i = \dfrac{A + \delta_m}{2}.

Step 3. Apply Snell at the first face, μ=sin⁡isin⁡r1\mu = \dfrac{\sin i}{\sin r_1}, with i=A+δm2i = \dfrac{A + \delta_m}{2} and r1=A2r_1 = \dfrac{A}{2}: μ=sin⁡ ⁣(A+δm2)sin⁡ ⁣(A2)\mu = \dfrac{\sin\!\big(\tfrac{A+\delta_m}{2}\big)}{\sin\!\big(\tfrac{A}{2}\big)}. That is equation (1).

Step 4. Use it both ways. Forward, with A=60∘A = 60^\circ, δm=40∘\delta_m = 40^\circ: μ=sin⁡50∘sin⁡30∘=0.7660.5=1.532\mu = \dfrac{\sin 50^\circ}{\sin 30^\circ} = \dfrac{0.766}{0.5} = 1.532. Backward, with A=60∘A = 60^\circ, μ=1.5\mu = 1.5: solve sin⁡ ⁣(60+δm2)=1.5×sin⁡30∘=0.75\sin\!\big(\tfrac{60 + \delta_m}{2}\big) = 1.5 \times \sin 30^\circ = 0.75, so 60+δm2=sin⁡−1(0.75)=48.59∘\tfrac{60 + \delta_m}{2} = \sin^{-1}(0.75) = 48.59^\circ, giving δm=2(48.59)−60=37.2∘\delta_m = 2(48.59) - 60 = 37.2^\circ, with the angle of incidence at that minimum i=48.59∘i = 48.59^\circ and internal angle r=30∘r = 30^\circ. Verified: μ = sin50/sin30 = 1.532, and δ_m = 37.2° (with i = 48.59°) from A = 60°, μ = 1.5, are recomputed in verify_hs2-prism-problems.py.

Go deeper · Derivation B · 1 minDesigning a two-prism combination — direct-vision and achromatic
Derivation BDesigning a two-prism combination — direct-vision and achromatic

The goal is to design a crown-flint pair for the two canonical outcomes: a direct-vision combination that spreads colour with no net bend, and an achromatic combination that bends with no net colour — landing A′=4.77∘A' = 4.77^\circ and A′=2.7∘A' = 2.7^\circ for a crown of A=6∘A = 6^\circ. The strategy is to write the net deviation and net dispersion of the two thin prisms and set the appropriate one to zero, solving for the unknown flint angle.

Step 1. Write the two thin-prism quantities. For the crown (AA) the mean deviation is (μy−1)A(\mu_y - 1)A and the dispersion is (μv−μr)A(\mu_v - \mu_r)A; for the flint (A′A', inverted) they are (μy′−1)A′(\mu'_y - 1)A' and (μv′−μr′)A′(\mu'_v - \mu'_r)A', subtracting from the crown's.

Step 2. Direct-vision: demand zero net deviation, (μy−1)A=(μy′−1)A′(\mu_y - 1)A = (\mu'_y - 1)A', so A′=(μy−1)Aμy′−1A' = \dfrac{(\mu_y - 1)A}{\mu'_y - 1}. With crown μy=1.517\mu_y = 1.517, A=6∘A = 6^\circ, flint μy′=1.65\mu'_y = 1.65: A′=0.517×60.65=3.1020.65=4.77∘A' = \dfrac{0.517 \times 6}{0.65} = \dfrac{3.102}{0.65} = 4.77^\circ. The colours still spread (the dispersions do not cancel), so you see the spectrum straight ahead.

Step 3. Achromatic: demand zero net dispersion, (μv−μr)A=(μv′−μr′)A′(\mu_v - \mu_r)A = (\mu'_v - \mu'_r)A', so A′=(μv−μr)Aμv′−μr′A' = \dfrac{(\mu_v - \mu_r)A}{\mu'_v - \mu'_r}. With crown μv−μr=1.523−1.514=0.009\mu_v - \mu_r = 1.523 - 1.514 = 0.009, A=6∘A = 6^\circ, and flint μv′−μr′=0.020\mu'_v - \mu'_r = 0.020: A′=0.009×60.020=0.0540.020=2.7∘A' = \dfrac{0.009 \times 6}{0.020} = \dfrac{0.054}{0.020} = 2.7^\circ. Now the colours recombine (a net mean deviation survives), which is exactly what an achromatic lens needs.

Step 4. Note the pattern. Both designs are one balance solved for A′A'; the only choice is which quantity to zero — deviation for direct-vision, dispersion for achromatic. The crown's dispersive power along the way is ω=(μv−μr)/(μy−1)=0.009/0.517=0.0174\omega = (\mu_v - \mu_r)/(\mu_y - 1) = 0.009/0.517 = 0.0174, and the crown's angular dispersion for A=6∘A = 6^\circ is (μv−μr)A=0.054∘(\mu_v - \mu_r)A = 0.054^\circ. Verified: A' = 4.77° (direct-vision), A' = 2.7° (achromatic), ω = 0.0174 and angular dispersion 0.054° are recomputed in verify_hs2-prism-problems.py.

Limiting cases

CaseResultMeaning
symmetric passageδ=δm\delta = \delta_m, r1=r2=A/2r_1 = r_2 = A/2the minimum, where equation (1) applies
thin prism (small AA)δ=(μ−1)A\delta = (\mu - 1)Adeviation linear in apex, independent of incidence
δm=A\delta_m = A (here 60∘60^\circ)μ=sin⁡A/sin⁡(A/2)=2cos⁡(A/2)\mu = \sin A/\sin(A/2) = 2\cos(A/2)for A=60∘A = 60^\circ, μ=3=1.732\mu = \sqrt{3} = 1.732 — a high-index limit
apex A>2CA > 2Cno emergent rayinternal ray exceeds the critical angle at the 2nd face — TIR
μ=1.5\mu = 1.5, so C=41.8∘C = 41.8^\circA≤83.6∘A \le 83.6^\circthe widest apex a crown prism transmits
dispersive powerω=(μv−μr)/(μy−1)\omega = (\mu_v - \mu_r)/(\mu_y - 1)independent of AA — a property of the glass alone

Worked examples

1 · Board level — refractive index from minimum deviation (CBSE and NEET pattern)

A prism of refracting angle 60∘60^\circ produces a minimum deviation of 40∘40^\circ. Find the refractive index of its glass.

Step 1. Recognise the type: given AA and δm\delta_m, asked for μ\mu — this is equation (1) used forward.

Step 2. Substitute: μ=sin⁡ ⁣(60+402)sin⁡ ⁣(602)=sin⁡50∘sin⁡30∘\mu = \dfrac{\sin\!\big(\tfrac{60 + 40}{2}\big)}{\sin\!\big(\tfrac{60}{2}\big)} = \dfrac{\sin 50^\circ}{\sin 30^\circ}.

Step 3. Evaluate: sin⁡50∘=0.766\sin 50^\circ = 0.766, sin⁡30∘=0.5\sin 30^\circ = 0.5, so μ=0.7660.5=1.532\mu = \dfrac{0.766}{0.5} = \boxed{1.532}. A typical dense-crown or light-flint glass.

Step 4. Sanity check: μ\mu should exceed 11 (glass is denser than air) and lie in the 1.41.4–1.71.7 range for common glasses — 1.531.53 sits comfortably there. Had you written sin⁡(40/2)\sin(40/2) on top you would have got sin⁡20∘/sin⁡30∘=0.684\sin 20^\circ/\sin 30^\circ = 0.684, a "refractive index" below one — the instant tell of the half-sum error.

2 · Board level — minimum deviation and a thin prism of crown glass (CBSE and ICSE pattern)

(a) Find the minimum deviation of a 60∘60^\circ crown-glass prism of refractive index 1.51.5, and the angle of incidence at which it occurs. (b) Find the deviation produced by a thin 5∘5^\circ crown prism of the same glass.

Step 1. Part (a), invert equation (1) for δm\delta_m: sin⁡ ⁣(60+δm2)=μ sin⁡ ⁣(602)=1.5×sin⁡30∘=1.5×0.5=0.75\sin\!\big(\tfrac{60 + \delta_m}{2}\big) = \mu\,\sin\!\big(\tfrac{60}{2}\big) = 1.5 \times \sin 30^\circ = 1.5 \times 0.5 = 0.75.

Step 2. Take the inverse sine: 60+δm2=sin⁡−1(0.75)=48.59∘\dfrac{60 + \delta_m}{2} = \sin^{-1}(0.75) = 48.59^\circ, so 60+δm=97.18∘60 + \delta_m = 97.18^\circ and δm=37.2∘\delta_m = \boxed{37.2^\circ}. The angle of incidence at minimum deviation is i=A+δm2=48.59∘i = \dfrac{A + \delta_m}{2} = 48.59^\circ, and the internal angle is r=A/2=30∘r = A/2 = 30^\circ.

Step 3. Part (b), a thin prism, so use equation (2): δ=(μ−1)A=(1.5−1)×5∘=0.5×5=2.5∘\delta = (\mu - 1)A = (1.5 - 1)\times 5^\circ = 0.5 \times 5 = \boxed{2.5^\circ}.

Step 4. Note the two regimes side by side: the 60∘60^\circ prism needs the full trigonometric formula (its angles are large), while the 5∘5^\circ prism uses the linear shortcut — and you must never apply (μ−1)A(\mu - 1)A to the 60∘60^\circ prism, which would give a wrong 30∘30^\circ instead of 37.2∘37.2^\circ.

3 · Competitive level — a direct-vision combination (JEE-Main and NEET pattern)

A crown-glass prism of refracting angle 6∘6^\circ and mean index 1.5171.517 is combined with a flint prism of mean index 1.651.65 so that the combination produces dispersion but no net mean deviation (a direct-vision prism). Find the flint prism's refracting angle.

Step 1. Direct-vision means zero net mean deviation, equation (5): (μy−1)A=(μy′−1)A′(\mu_y - 1)A = (\mu'_y - 1)A', the crown's deviation balanced by the flint's.

Step 2. Solve for the flint angle: A′=(μy−1)Aμy′−1=(1.517−1)×61.65−1=0.517×60.65A' = \dfrac{(\mu_y - 1)A}{\mu'_y - 1} = \dfrac{(1.517 - 1)\times 6}{1.65 - 1} = \dfrac{0.517 \times 6}{0.65}.

Step 3. Evaluate: A′=3.1020.65=4.77∘A' = \dfrac{3.102}{0.65} = \boxed{4.77^\circ}. The flint prism is smaller than the crown because flint bends more strongly per degree of apex.

Step 4. Confirm the colour survives: the two dispersions do not cancel (they would need equation (6), a different balance), so white light still fans into a spectrum — seen straight ahead, which is the whole purpose of a direct-vision prism in a pocket spectroscope. Timing note: under two minutes. The only decision is which balance to use — deviations for direct-vision, dispersions for achromatic — and once chosen it is a single division. Write the balance equation before you touch the calculator.

4 · Competitive level — dispersive power and an achromatic pair (JEE-Advanced and NEET pattern)

Crown glass has μr=1.514\mu_r = 1.514, μy=1.517\mu_y = 1.517, μv=1.523\mu_v = 1.523. (a) Find its dispersive power. (b) A 6∘6^\circ crown prism is to be made achromatic (no net dispersion) with a flint prism whose (μv′−μr′)=0.020(\mu'_v - \mu'_r) = 0.020. Find the flint prism's refracting angle.

Step 1. Part (a), dispersive power, equation (4): ω=μv−μrμy−1=1.523−1.5141.517−1=0.0090.517=0.0174\omega = \dfrac{\mu_v - \mu_r}{\mu_y - 1} = \dfrac{1.523 - 1.514}{1.517 - 1} = \dfrac{0.009}{0.517} = \boxed{0.0174} — small and dimensionless, a property of crown glass alone.

Step 2. Part (b), achromatic means zero net dispersion, equation (6): (μv−μr)A=(μv′−μr′)A′(\mu_v - \mu_r)A = (\mu'_v - \mu'_r)A', the crown's dispersion balanced by the flint's.

Step 3. Solve: A′=(μv−μr)Aμv′−μr′=0.009×60.020=0.0540.020=2.7∘A' = \dfrac{(\mu_v - \mu_r)A}{\mu'_v - \mu'_r} = \dfrac{0.009 \times 6}{0.020} = \dfrac{0.054}{0.020} = \boxed{2.7^\circ}.

Step 4. Read the physics: because the flint disperses more strongly, a smaller flint prism (2.7∘2.7^\circ) cancels the crown's colour spread — but leaves a net deviation (the mean beam is bent), which is exactly what makes an achromatic combination useful in a lens: colours focus together, and the beam is still steered. The crown's own angular dispersion here is (μv−μr)A=0.009×6=0.054∘(\mu_v - \mu_r)A = 0.009 \times 6 = 0.054^\circ. Timing note: about two minutes for both parts. The dispersive power is one subtraction over another; the achromat is one balance solved for A′A'. Keep crown and flint quantities in separate columns so you never balance a crown deviation against a flint dispersion.

Exam traps

The half-sum over the half-apex. Equation (1) is sin⁡ ⁣(A+δm2)\sin\!\big(\tfrac{A+\delta_m}{2}\big) over sin⁡ ⁣(A2)\sin\!\big(\tfrac{A}{2}\big). Writing sin⁡(δm/2)\sin(\delta_m/2), or failing to halve, is the number-one lost mark — it produces a "refractive index" below one. Anchor it: at minimum deviation the incidence angle is (A+δm)/2(A+\delta_m)/2 and the internal angle is A/2A/2.

Thin-prism formula on a fat prism. δ=(μ−1)A\delta = (\mu - 1)A holds only for small AA. Apply it to a 60∘60^\circ prism and you get 30∘30^\circ instead of the correct 37.2∘37.2^\circ. Use the full formula whenever AA is more than a few degrees.

Balancing the wrong quantity in a combination. Direct-vision (no net bend) balances the deviations (μy−1)A(\mu_y - 1)A; achromatic (no net colour) balances the dispersions (μv−μr)A(\mu_v - \mu_r)A. Swapping the two gives a plausible-looking but wrong flint angle.

Getting the colour order backwards. Violet has the higher index, so violet deviates more than red. A spectrum drawn with red bent most is wrong, and any "μv<μr\mu_v < \mu_r" is a red flag.

Ignoring the apex limit. A prism with A>2CA > 2C transmits nothing — the internal ray total-internally-reflects at the second face. For μ=1.5\mu = 1.5, C=41.8∘C = 41.8^\circ, so AA must stay below 83.6∘83.6^\circ. Always sanity-check before quoting a deviation.

Board scope. Prism deviation-and-dispersion numericals are examined by CBSE, ICSE, JEE-Main, JEE-Advanced and NEET. IGCSE 0625 treats dispersion by a prism only qualitatively (white light into the seven colours) and sets no minimum-deviation or dispersive-power calculations. AP Physics 2 carries none of this, and IB and A-Level have no prism optics. If your board is one of those, this gym is background — learn the qualitative spectrum instead.

Origins · 1 minIsaac Newton and how this was discovered
IN
Isaac Newton
1642–1727 · Cambridge

Newton is the reason the prism problem exists. In 1666, in the plague years away from Cambridge, he let a sunbeam through a prism and found not a smear but a fan — an oblong spectrum of pure colours — and then did the decisive experiment (his experimentum crucis): he passed a single colour through a second prism and it bent, but did not split further. White light, he concluded, is not modified by the prism but separated: it is a mixture, and the prism disperses it because each colour is refracted by a different amount. He reported this to the Royal Society in 1672 and laid it out fully in Opticks (1704), one of the most influential books in science. The minimum-deviation formula you compute with is later, Snell-based work, and the dispersive-power framework was refined by Fraunhofer's precise measurements a century on — but the idea that a prism deviates and disperses, and that the deviation depends on colour, is Newton's, straight from that sunbeam.

History
1666
— Isaac Newton disperses sunlight with a prism into the spectrum.
1672
— Newton reports his experimentum crucis on the composition of light to the Royal Society.
1704
— Newton publishes Opticks, the study of light and colour.
1733
— Chester Moore Hall builds the first achromatic lens, combining crown and flint glass.
1758
— John Dollond patents the achromatic doublet.
1814
— Joseph von Fraunhofer maps the dark solar lines, fixing standard dispersion wavelengths.

Going further

Texts. NCERT Physics Class 12, Ch. 9 ("Ray Optics and Optical Instruments"), refraction through a prism, dispersion and the dispersive-power combinations · H. C. Verma, Concepts of Physics Vol. 1, "Dispersion and Spectra" · D. C. Pandey, Optics and Modern Physics, the prism problem set for JEE and NEET.

Provenance. Standard optics. Recompute: μ=sin⁡50∘/sin⁡30∘=1.532\mu = \sin 50^\circ/\sin 30^\circ = 1.532; δm=37.2∘\delta_m = 37.2^\circ (with i=48.59∘i = 48.59^\circ, r=30∘r = 30^\circ) from A=60∘A = 60^\circ, μ=1.5\mu = 1.5; thin-prism δ=(1.5−1)×5=2.5∘\delta = (1.5 - 1)\times 5 = 2.5^\circ; dispersive power ω=0.009/0.517=0.0174\omega = 0.009/0.517 = 0.0174; angular dispersion (μv−μr)×6=0.054∘(\mu_v - \mu_r)\times 6 = 0.054^\circ; direct-vision A′=0.517×6/0.65=4.77∘A' = 0.517 \times 6/0.65 = 4.77^\circ; achromatic A′=0.009×6/0.020=2.7∘A' = 0.009 \times 6/0.020 = 2.7^\circ; the grazing-limit index μ=3=1.732\mu = \sqrt{3} = 1.732 when δm=A=60∘\delta_m = A = 60^\circ; and the apex limit A≤2C=2sin⁡−1(1/1.5)=83.6∘A \le 2C = 2\sin^{-1}(1/1.5) = 83.6^\circ.

Prerequisites. The Prism: Deviation, Dispersion and Dispersive Power — the concept article that derives these relations, which this gym drills. Snell's law and the critical angle were established earlier in this chapter and are quoted here. Leads to. Instrument Magnification Numericals — the sibling gym for the optical instruments · the university Dispersion by Prisms seam.

Go deeper (UG). The wavelength-dependence of refractive index itself — the Cauchy and Sellmeier dispersion formulas that give μ(λ)\mu(\lambda), and the physics of why glass disperses — is developed in Dispersion by Prisms and Chromatic Aberration and Achromats in the university optics track. Nothing here needs it; the prism relations are pure ray geometry plus a quoted Snell's law.

Why is the deviation a minimum exactly at symmetric passage?

Because the deviation depends smoothly on the angle of incidence and the two prism faces are geometrically interchangeable, so the deviation you get entering at angle ii and leaving at ee is the same as entering at ee and leaving at ii. A smooth function that is symmetric under swapping ii and ee has its turning point where i=ei = e — which is exactly the symmetric passage, with r1=r2=A/2r_1 = r_2 = A/2. Physically, as you rotate the prism the deviation falls, flattens at this symmetric point, then rises again, tracing a shallow U. The flatness at the bottom is what makes the measurement precise: near minimum deviation the angle barely changes with small rotations, so you can pin δm\delta_m down cleanly and get μ\mu from equation (1).

When can I use the thin-prism formula and when must I use the full one?

Use δ=(μ−1)A\delta = (\mu - 1)A only when the refracting angle is small — a few degrees, up to perhaps 10∘10^\circ — because it comes from replacing every sine by its angle, which is only accurate for small angles. For a 5∘5^\circ or 6∘6^\circ prism it is excellent and saves all the trigonometry. For a 60∘60^\circ prism it is simply wrong: it would give (1.5−1)×60=30∘(1.5 - 1)\times 60 = 30^\circ, whereas the true minimum deviation is 37.2∘37.2^\circ. The rule of thumb: if the problem says "thin prism" or gives an apex of a few degrees, use the linear formula; if it gives a 30∘30^\circ, 45∘45^\circ or 60∘60^\circ prism, use the full μ=sin⁡ ⁣(A+δm2)/sin⁡ ⁣(A2)\mu = \sin\!\big(\tfrac{A+\delta_m}{2}\big)/\sin\!\big(\tfrac{A}{2}\big).

What is the difference between angular dispersion and dispersive power?

Angular dispersion is an angle: the actual spread between the violet and red rays, (μv−μr)A(\mu_v - \mu_r)A, and it depends on the prism's apex AA — a bigger prism spreads the colours more. Dispersive power is a ratio: the angular dispersion divided by the mean deviation, ω=(μv−μr)/(μy−1)\omega = (\mu_v - \mu_r)/(\mu_y - 1), and the apex cancels, so it is a pure number characterising the glass itself, not the prism. Think of angular dispersion as "how much this particular prism spreads the colours" and dispersive power as "how much colour spread this kind of glass gives per unit of bending." Crown glass has low dispersive power (about 0.0170.017), flint glass higher (about 0.030.03–0.040.04), and that difference is exactly what lets you combine them to cancel colour while keeping deviation.

Why does violet deviate more than red?

Because a glass's refractive index is slightly larger for violet light than for red — the glass is "denser" to short wavelengths. Deviation grows with index (more index, more bending), so violet, with the higher μ\mu, is bent the most and red the least, with the other colours in between. This is dispersion, and it is why white light fans into a spectrum with red at the top of the fan (least deviated) and violet at the bottom (most deviated). The fact that μ\mu depends on colour at all — called dispersion of the medium — is the whole reason prisms make spectra; if μ\mu were the same for every colour, a prism would deviate white light but never split it.

What happens if the prism angle is too big?

The ray never comes out. As the apex AA grows, the internal angle r2r_2 at the second face grows too, and once it exceeds the critical angle CC the ray total-internally-reflects at that face instead of emerging. The limit is A≤2CA \le 2C: for crown glass with μ=1.5\mu = 1.5, the critical angle is C=sin⁡−1(1/1.5)=41.8∘C = \sin^{-1}(1/1.5) = 41.8^\circ, so the widest apex that still transmits is about 83.6∘83.6^\circ. Push past that and there is no emergent ray and no deviation to speak of — which is precisely how a right-angle prism is used as a perfect mirror in binoculars and periscopes. So in any prism problem, a quick check that A<2CA < 2C confirms the light actually gets through before you quote a deviation.

Do the combination formulas assume the prisms are thin?

Yes — the direct-vision and achromatic balances, equations (5) and (6), are written with the thin-prism deviation (μ−1)A(\mu - 1)A and dispersion (μv−μr)A(\mu_v - \mu_r)A, which hold only for small refracting angles. That is fine, because combination prisms (in spectroscopes and achromatic lenses) are deliberately made thin, with apex angles of a few degrees, so the linear formulas apply and the algebra stays simple. The design is then just one balance solved for the unknown flint angle. For large-angle prisms you would have to use the full deviation for each colour separately and balance numerically, which is beyond the scope of these problems — and beyond what any of these boards asks. Whenever you see a combination question, treat both prisms as thin unless told otherwise.

Before this

Where this leads