Visual walkthrough — Thermal protection — silica tiles (Shuttle), UHTCs (ZrB₂, HfB₂)

---

Step 1 — The picture: a fast vehicle plowing into still air

WHAT. We set up the scene: a vehicle moving at speed $V$, air of density $\rho$ ("how many kilograms of air sit in each cubic metre") sitting ahead of it.

WHY. Before any formula, we need to know what is actually happening physically — energy is being transferred from your motion into the air. Everything downstream is bookkeeping on that one fact.

PICTURE — Figure s01. The red vehicle sweeps out a tube of air per second. That swept air is what gets heated.

Figure s01 — vehicle at speed $V$ meeting still air of density $\rho$.

---

Step 2 — How much air do you hit per second? (mass flux)

WHAT. We compute the rate at which air-mass arrives at the vehicle, called the mass flux $\dot{m}$.

WHY. Heat comes from the air you disturb. To find heating we first must count how much air per second you are disturbing. You cannot heat air you never touch.

WHY this tool (a rate, the little dot). The dot on $\dot{m}$ means "per second" — a rate. We use a rate, not a total, because re-entry heating is about power (energy per second), so every quantity must be "per second" to match.

$\underset{\begin{array}{c}\text{kg of air}\\ \text{hit per second}\end{array}}{\underbrace{\dot{m}}}=\underset{\begin{array}{c}\text{air}\\ \text{density}\end{array}}{\underbrace{\rho }}\cdot \underset{\begin{array}{c}\text{front}\\ \text{area}\end{array}}{\underbrace{A}}\cdot \underset{\begin{array}{c}\text{speed}\end{array}}{\underbrace{V}}$

PICTURE — Figure s02. The shaded box is the air swept in one second: length $V$, face $A$.

Figure s02 — swept-air box gives $\dot{m}=\rho AV$.

---

Step 3 — How much energy does each kilogram of air carry? ($V^2$)

WHAT. We find the energy per kilogram of incoming air.

WHY. Mass alone is not heat. Each kilogram carries a punch of energy that gets dumped into the shock layer. We need energy-per-kilogram so we can multiply it by kilograms-per-second later.

WHY this tool (kinetic energy $\frac{1}{2}m{V}^2$). Kinetic energy is the quantity that measures "how much motion-energy a moving mass has." The air's motion is exactly what gets converted to heat, so this is the right ledger. The important feature is the square: energy grows as $V^2$, not as $V$.

$\underset{\text{joules per kg}}{\underbrace{\text{energy per kg}}}=\frac{1}{2}\underset{\begin{array}{c}\text{speed}\\ \text{squared}\end{array}}{\underbrace{V^2}}\propto V^2$

The $\frac{1}{2}$ is a numerical constant. From here on we care only about how the energy grows with speed, so we drop the $\frac{1}{2}$ and keep just the scaling $\propto V^2$ — this is the first place a $=$ becomes a $\propto$.

PICTURE — Figure s03. Two parabolas: at double the speed, the energy bar is four times as tall.

Figure s03 — energy per kg grows as $V^2$; double $V$ → 4× energy.

---

Step 4 — Multiply: the heat flux gets its $V^3$

WHAT. We combine Steps 2 and 3 into the heat flux $q$ (power per unit area, $W/{\mathrm{m}}^2$).

WHY. This multiplication is the heart of the whole formula. It is where the famous $V^3$ is born, and it is pure bookkeeping — no advanced physics needed.

$q\sim \underset{\begin{array}{c}\text{mass flux}\\ \text{per area}\\ (\text{Step 2})\end{array}}{\underbrace{(\rho V)}}\times \underset{\begin{array}{c}\text{energy per kg}\\ (\text{Step 3})\end{array}}{\underbrace{V^2}}\propto \rho \underset{=V^3}{\underbrace{V\cdot V^2}}=\rho V^3$

(The $\sim$ signals that the $\frac{1}{2}$ from kinetic energy has been dropped — we track only the scaling.)

• The $\rho$ came from how much air (Step 2).
• One power of $V$ came from how fast the air arrives (Step 2).
• Two powers of $V$ came from the energy each kg carries (Step 3).
• Together: $V\times V^2=V^3$.

PICTURE — Figure s04. A staircase: each factor stacks a power of $V$, ending at the tall $V^3$ column.

Figure s04 — stacking powers of $V$: $\rho \cdot V\cdot V^2=\rho V^3$.

---

Step 5 — The cushion of slowed air (why a boundary layer even exists)

WHAT. We introduce two new symbols: the boundary-layer thickness $\delta$ and the viscosity $\nu$ (Greek "nu"), a measure of how sticky the air is (units ${\mathrm{m}}^2/s$).

WHY. Steps 1–4 counted the total energy available in the shock. But how much of it actually gets to the wall is throttled by this cushion. Without viscosity there would be no cushion at all, so $\nu$ must appear.

WHY this tool (viscosity balancing convection). The cushion's thickness is set by a tug-of-war: sticky friction ($\nu$) tries to grow the slow layer outward, while the fast oncoming flow ($V$) sweeps it back thin. Near a nose of radius $R_n$, the flow spends a "residence time" of about $R_n/V$ curving around the tip; in that time, viscous drag diffuses inward a distance $\delta \sim \sqrt{\nu \times \text{time}}=\sqrt{\nu R_n/V}$. This $\sqrt{\text{diffusion}}$ form is the same square-root growth you get whenever something spreads by diffusion in a given time.

$\underset{\begin{array}{c}\text{cushion}\\ \text{thickness}\end{array}}{\underbrace{\delta }}\sim \sqrt{\underset{\begin{array}{c}\text{air}\\ \text{stickiness}\end{array}}{\underbrace{\nu }}\cdot \underset{\begin{array}{c}\text{residence}\\ \text{time at nose}\end{array}}{\underbrace{\frac{R_n}{V}}}}=\sqrt{\frac{\nu R_n}{V}}$

Read the three dependences off the picture:

• $\delta \uparrow$ with $\nu$: stickier air, thicker cushion.
• $\delta \uparrow$ with $R_n$: a fatter nose gives the flow longer to build the cushion.
• $\delta \downarrow$ with $V$: faster flow sweeps the cushion thinner. Note the $\delta \propto V^{-1/2}$ — remember this exponent, it is the piece that keeps the final answer at exactly $V^3$ in Step 6.

PICTURE — Figure s05. The red band is $\delta$, labelled and dimensioned; the nose radius $R_n$ and the free-stream $V$ are both drawn.

Figure s05 — viscous cushion of thickness $\delta \sim \sqrt{\nu R_n/V}$ around a nose of radius $R_n$.

---

Step 6 — Conduct heat across the cushion → the full scaling (with careful V-bookkeeping)

WHAT. Apply Fourier's law with the $\delta$ we found, keeping the full $V$-dependence of every factor (this is where the reviewer-worthy care lives), then fold in density.

WHY. This is the single step that (a) turns "sharp = hot" into $R_n^{-1/2}$, (b) drops the density exponent from $1$ to $\frac{1}{2}$, and (c) must reproduce exactly $V^3$, no more, no less. We now track every power of $V$ explicitly so nothing is hidden.

Sub-step 6a — Fourier across the cushion, every factor kept. $q_{\text{wall}}=\underset{\begin{array}{c}\text{air’s thermal}\\ \text{conductivity}\end{array}}{\underbrace{k_{\text{air}}}}\frac{\stackrel{\begin{array}{c}\text{temp drop}\\ \propto V^2(\text{Step 3})\end{array}}{\overbrace{\Delta T}}}{\underset{\begin{array}{c}\text{cushion}\\ \propto V^{-1/2}(\text{Step 5})\end{array}}{\underbrace{\delta }}}$ Two of these factors carry $V$:

• $\Delta T\propto V^2$ — the stagnation temperature rise is the kinetic energy per kg from Step 3.
• $\delta \propto V^{-1/2}$ — from Step 5; dividing by a shrinking cushion multiplies the flux, so $1/\delta \propto V^{+1/2}$.

Sub-step 6b — collect the $V$ exponents (the missing bookkeeping). $q_{\text{wall}}\propto \frac{\Delta T}{\delta }\propto \frac{V^2}{V^{-1/2}}=V^2\cdot V^{1/2}=V^{5/2}$ So conduction alone gives $V^{5/2}$, not $V^3$. We are still missing exactly half a power of $V$ — it comes from the density factor in 6c, because in a shock the compressed gas density itself rises with speed. This is the reconciliation the previous version skipped.

Sub-step 6c — the density factor and where ${\rho }^{1/2}$ and the last $V^{1/2}$ come from. Two density facts enter:

1. The gas conducting heat is shock-compressed, and across a strong shock the density behind it grows in proportion to the oncoming momentum, so the effective density feeding conduction rises with speed: it contributes one more $V^{1/2}$. That is the missing half-power: $V^{5/2}\times V^{1/2}=V^3$. ✓ Now the $V$ exponent is exactly 3, fully accounted.
2. The free-stream density $\rho$ enters at half power. Picture it in the cushion: a denser gas packs more heat-carrying molecules (pushes conduction up), but a denser gas also makes the cushion behave more sluggishly (pushes the effective diffusion down). These two opposite pulls partly cancel, so $\rho$ survives at ${\rho }^{1/2}$, not ${\rho }^1$ — exactly the Fay–Riddell result, now with a physical reason rather than a quoted number.

$\boxed{q\propto \underset{\begin{array}{c}\text{sharp = hot}\\ (\text{Step 5})\end{array}}{\underbrace{\frac{1}{\sqrt{R_n}}}}\underset{\begin{array}{c}\text{two pulls}\\ \text{cancel to }\sqrt{\rho }\end{array}}{\underbrace{{\rho }^{1/2}}}\underset{\begin{array}{c}{V}^{5/2}\text{ (conduction)}\\ +V^{1/2}\text{ (shock density)}\end{array}}{\underbrace{V^3}}}$

PICTURE — Figure s06. A ledger: the $V$-exponent tally ($V^2$ from $\Delta T$, $V^{1/2}$ from $1/\delta$, $V^{1/2}$ from shock density) summing to $V^3$, plus the two opposing $\rho$-arrows cancelling to $\sqrt{\rho }$.

Figure s06 — bookkeeping of $V$ exponents $2+\frac{1}{2}+\frac{1}{2}=3$, and the $\rho$-arrows cancelling to ${\rho }^{1/2}$.

---

Step 7 — The fork: block the heat vs take the heat

WHAT. We split the arriving heat into a radiated part and a conducted part.

WHY. The two philosophies of the whole chapter fall directly out of this balance. There is no third option: heat that arrives must leave, either as light out or as conduction in.

WHY this tool (Stefan–Boltzmann). A hot surface radiates power $\epsilon \sigma T^4$. The fourth power is decisive: get a surface hot enough and it dumps enormous power back to space. We use this law here — not conduction — because in steady state radiation is the only escape route pointing back outward; a hot surface has no other way to shed heat to empty space.

$\underset{\text{arrives}}{\underbrace{q}}=\underset{\begin{array}{c}\text{radiated}\\ \text{out (glow)}\end{array}}{\underbrace{\epsilon \sigma T^4}}+\underset{\begin{array}{c}\text{conducted}\\ \text{inward (solid)}\end{array}}{\underbrace{k\frac{\Delta T}{L}}}$

• $\epsilon$ (emissivity, 0–1) = how good a glower the surface is. Black glaze ≈ 0.9.
• $\sigma =5.67\times 10^{-8}\mathrm{W}{\mathrm{m}}^{-2}{\mathrm{K}}^{-4}$ = Stefan–Boltzmann constant.
• $T$ = surface temperature in kelvin.
• $k$ = thermal conductivity of the solid shield, $L$ = tile thickness, $\Delta T$ = temperature drop across it. (This $=$ is a real balance — all constants kept.)

Two design answers:

• Silica tile (block): make $k$ tiny so the conducted term is nearly zero. Let $\epsilon \sigma T^4$ carry almost all of $q$ — radiate it away, keep the inner face cold. Broad surfaces, ~1260 °C.
• UHTC (take): on a sharp edge $L\to 0$ (no volume to insulate), so you cannot rely on the conducted term being blocked. Instead pick a material whose $T$ from $q=\epsilon \sigma T^4$ is below its melting point — genuinely stable, self-healing oxide, high $k$ to spread heat. Sharp edges, >1700 °C.

PICTURE — Figure s07. Arriving red arrow splits: one arrow radiates back out ($T^4$), one leaks inward ($k\Delta T/L$).

Figure s07 — arriving heat $q$ splits into radiate-out $\epsilon \sigma T^4$ vs conduct-in $k\Delta T/L$.

---

Step 8 — Edge & limiting cases (never leave the reader stranded)

Case A — Standing still, $V=0$. Then $q\propto V^3=0$. No motion, no swept air, no heat. Sanity ✓.

Case B — Perfect insulator, $k\to 0$. In Step 7 the conducted term $k\Delta T/L\to 0$, so all of $q$ must radiate: $q=\epsilon \sigma T^4$. The inner face is fully protected — this is the silica-tile ideal.

Case C — Perfectly sharp nose, $R_n\to 0$. From Step 5, $\delta \sim \sqrt{\nu R_n/V}\to 0$: the cushion vanishes. Then $q\propto R_n^{-1/2}\to \infty$. Heat flux blows up. This is why a truly infinitely-sharp edge is impossible to cool by insulation — you must take the heat with a UHTC, and even then $R_n$ is kept small-but-finite.

Case D — The velocity ratio (the Moon-return number, promised in Step 4). With $\rho ,R_n,\nu$ comparable, only $V^3$ differs: $\frac{q_{\text{Moon}}}{q_{\text{LEO}}}={\left(\frac{V_{\text{Moon}}}{V_{\text{LEO}}}\right)}^3={\left(\frac{11}{7.8}\right)}^3\approx 2.8$ About 3× worse — matching why Apollo used an ablative shield, not reusable tiles.

PICTURE — Figure s08. Three mini-panels: $V=0$ flat line; $R_n\to 0$ spike to infinity; the $V^3$ curve with LEO and Moon marked.

Figure s08 — edge cases: $V=0$ gives zero, $R_n\to 0$ blows up, and the $V^3$ curve with LEO vs Moon marked.

---

The one-picture summary

Figure s09 — the whole derivation in one frame: $\rho V\times V^2\to \rho V^3$ shock energy; conduction throttle across $\delta$ reallocates to ${\rho }^{1/2}{V}^3/\sqrt{R_n}$; the arriving heat forks into radiate-out vs conduct-in — silica vs UHTC.

The whole derivation in a single frame: swept air ($\rho V$) × kinetic energy ($V^2$) → $\rho V^3$ shock energy; conducting across the viscous cushion $\delta \sim \sqrt{\nu R_n/V}$ throttles this to the parent formula ${\rho }^{1/2}{V}^3/\sqrt{R_n}$ (with $V$ exponents $2+\frac{1}{2}+\frac{1}{2}=3$ bookkept in Step 6); the arriving heat then splits into radiate-out ($\epsilon \sigma T^4$) vs conduct-in ($k\Delta T/L$), and that split is the fork between silica (block) and UHTC (take).

Recall Feynman retelling — say it in plain words

You're a spacecraft flying into still air. First, meet the star of the show: $q$, the heat flux — how hard each square metre of your skin is being roasted, in watts per square metre. Everything is a hunt for what makes $q$ big. Each second you sweep out a tube of air: how much you hit is density times speed ($\rho V$). Each kilogram of that air carries kinetic energy that grows with speed squared ($V^2$) — because energy of motion always scales with the square of speed; that same $V^2$ is also the temperature rise of the shock gas. Multiply "how much air" by "how much energy each carries" and you get heat per second: $\rho V\times V^2=\rho V^3$. That's the raw shock energy; double the speed and it's eight times bigger. But that heat can't touch the wall directly — air is sticky (viscosity $\nu$), so there's always a thin cushion of slowed air, the boundary layer, of thickness $\delta$. Near a nose of radius $R_n$ the flow curves around in a time about $R_n/V$, and stickiness diffuses inward $\sqrt{\nu R_n/V}$ in that time — that's $\delta$, and it thins like $V^{-1/2}$. A sharp nose (small $R_n$) grows a thinner cushion, so heat conducts through faster. Push Fourier's law across the cushion — flux is conductivity times temperature drop over thickness — and count the speeds honestly: $\Delta T$ gives $V^2$, the thin cushion gives another $V^{1/2}$, so conduction alone is $V^{5/2}$; the shock-compressed density supplies the last $V^{1/2}$ to make exactly $V^3$. The free-stream density enters only at half power, because a denser gas both carries more heat and diffuses more sluggishly and those fight to a draw — that's the ${\rho }^{1/2}$. Also remember: in a real shock the air is thousands of kelvin, so $k_{\text{air}}$ and $\nu$ aren't really constants — that's why engineers use the full correlation. Put it together and you get the parent formula ${\rho }^{1/2}{V}^3/\sqrt{R_n}$. Once heat lands on the surface it has two exits: glow it back to space (radiation, $\propto T^4$, incredibly powerful when hot) or leak it inward (solid conduction, $k\Delta T/L$). Make the material a super-insulator with a black glowing coat and it radiates almost everything away and stays cold inside — that's a silica tile, "block the heat." A razor-sharp edge has no room to insulate, so instead pick a ceramic that's genuinely stable at that glowing temperature, spreads heat with high conductivity, and grows a self-healing glass skin — that's a UHTC, "take the heat." The single number that decides which one you need is $T=(q/\epsilon \sigma {)}^{1/4}$: if it's above silica's limit, you're forced into a UHTC.

Recall Quick self-test (cover the answers)

Where does the $V^3$ come from? ::: Mass flux $\rho V$ (Step 2) times kinetic energy per kg $\propto V^2$ (Step 3); the conduction step re-splits it as $V^2$ from $\Delta T$ + $V^{1/2}$ from $1/\delta$ + $V^{1/2}$ from shock density = $V^3$. What does the symbol $q$ mean? ::: Heat flux — energy hitting each square metre of surface per second, in $W/{\mathrm{m}}^2$. Why does a boundary-layer cushion exist at all? ::: Air has viscosity $\nu$ (stickiness), so the layer touching the wall is dragged to a near-stop. What sets the cushion thickness $\delta$? ::: $\delta \sim \sqrt{\nu R_n/V}$ — viscosity times residence time $R_n/V$, a diffusion square-root; it thins as $V^{-1/2}$. Why is a sharp nose hotter? ::: Smaller $R_n$ → thinner cushion $\delta$ → steeper temperature gradient in Fourier's law → $q\propto 1/\sqrt{R_n}$. Why does density enter at ${\rho }^{1/2}$, not $\rho$? ::: A denser gas both carries more heat and diffuses more sluggishly; the two opposing effects partly cancel, leaving $\sqrt{\rho }$ (Fay–Riddell). Are $k_{\text{air}}$ and $\nu$ really constants in a shock? ::: No — the gas hits thousands of kelvin and both rise steeply with temperature; the scaling law absorbs this into its "constants." What are the only two exits for arriving heat? ::: Radiate out ($\epsilon \sigma T^4$) or conduct into the solid ($k\Delta T/L$). How much worse is lunar return than LEO? ::: $(11/7.8{)}^3\approx 2.8$, about 3×.