Thermal Expansion of Solids & Liquids

Thermal expansion is engineered around everywhere you look. Bridges and elevated roads sit on expansion joints — those finger-like steel combs and rubber gaps in the deck — so a hot afternoon's growth is absorbed instead of buckling the span. Long railway lines are either laid with gaps or pre-stressed and clamped so that summer expansion cannot kink the track. Overhead power lines are strung with deliberate slack because they sag lower on hot days. In the workshop, engineers exploit expansion on purpose: shrink-fitting slips a heated steel ring over a shaft and lets it grip as it cools, and rivets are hammered in hot so they clamp tight while contracting. Bimetallic strips — two metals of different $\alpha$ bonded together — bend with temperature to switch thermostats, irons and old car indicators on and off, the direct subject of the thermal-stress and bimetallic archetype. Even the humble glass-and-mercury (or coloured-alcohol) thermometer is nothing but equation (4) in a tube. And the same falling-density-with-warmth from equation (5) is what lifts a hot-air balloon and stirs every convection current in the atmosphere and oceans.

Zoom into a crystal and you find atoms tethered to their neighbours by bonds that behave like tiny springs, jiggling about fixed average positions. Raising the temperature pumps energy into that jiggling, so each atom swings through a wider arc. If the bond were a perfect spring — a symmetric parabola, equally hard to stretch and to squeeze — the average position would not move at all, and matter would not expand. It expands because the bond is asymmetric: the atoms resist being pushed together far more stiffly than they resist being pulled apart. Give an atom a wider swing and it spends more of its time on the easy, stretched side than on the hard, squeezed side, so its average separation slides outward. Expansion is that outward slide, summed over every bond in the body.

This is why almost every material expands on heating, why the expansion is small (bonds are stiff), and why it is very nearly proportional to the temperature change over the modest ranges a school problem uses. It also quietly warns us that the neat linear law is an approximation: over hundreds of degrees the coefficient itself drifts, and a few oddballs — water between 0 and 4 °C, the star exception — even contract. For ordinary solids over ordinary ranges, though, one clean rule captures it all, and the same rule reappears in three geometric costumes.

Take a rod of original length $L_0$ and warm it through a temperature change $\Delta T$. Experiment — and the molecular picture above — say the extra length is proportional both to how much rod there is and to how hard you heat it:

where: $L_0$ — original length (any unit; the answer comes out in the same unit); $\Delta T$ — temperature change (a rise, identical in °C and K, since a difference of one degree is the same on both scales); $\alpha$ — the coefficient of linear expansion, a property of the material with units $\mathrm{K^{-1}}$ (per kelvin). Read it physically: $\alpha$ is the fractional stretch per degree, $\alpha = (\Delta L / L_0)/\Delta T$, so a steel value $\alpha_{\text{steel}} = 1.2\times10^{-5}\ \mathrm{K^{-1}}$ means each metre of steel lengthens by twelve micrometres for every one-degree rise. The numbers are tiny — which is exactly why expansion sneaks up on engineers only at large sizes or large temperature swings. Two features of equation (1) earn marks by themselves. First, the fractional change $\Delta L/L_0 = \alpha\,\Delta T$ depends only on the material and the temperature change, never on the length — a short steel ruler and a long steel bridge stretch by the same percentage. Second, a hole or cavity expands exactly as if it were filled with the surrounding metal: heat a steel washer and the hole gets bigger, not smaller, because every point of the metal moves outward from every other. That ring-and-ball demonstration is a NEET and IGCSE favourite precisely because intuition guesses the opposite.

A flat plate has two dimensions and a solid block has three, but they are made of the same expanding material, so their coefficients are not new constants to memorise — they are forced by geometry. Warm a square plate and each side grows by the factor $1+\alpha\,\Delta T$; the area, being side times side, grows by that factor squared:

where: $A_0$ — original area; $\beta$ — the coefficient of area (superficial) expansion. Read it as "two lengths, so twice the coefficient." For steel, $\beta = 2\alpha_{\text{steel}} = 2.4\times10^{-5}\ \mathrm{K^{-1}}$, so a $1\ \mathrm{m^2}$ steel sheet heated by $40\ \mathrm{K}$ gains $\Delta A = 1\times2.4\times10^{-5}\times40 = 9.6\times10^{-4}\ \mathrm{m^2}$ — about $9.6\ \mathrm{cm^2}$, a barely visible border. A solid block has three dimensions, so its volume grows by the cube of the same factor:

where: $V_0$ — original volume; $\gamma$ — the coefficient of volume (cubical) expansion, three times the linear one. The single most-punished trap in this whole topic is writing $\beta = 3\alpha$ or $\gamma = 2\alpha$: area gets the 2, volume gets the 3. The next derivation shows exactly where the 2 and the 3 come from, so you can rebuild them under exam pressure instead of gambling on memory.

A liquid has no shape of its own — it takes the shape of its vessel — so it shows only volume expansion, described by a coefficient $\gamma$ just like a solid's. But there is a catch that examiners love: you can only watch a liquid expand relative to its container, and the container is expanding too. When you heat mercury in a glass tube, the mercury column rises by less than the mercury's true expansion, because the glass cavity holding it has grown as well. What you see is the difference:

where: $\gamma_{\text{real}}$ — the liquid's own (absolute) volume coefficient; $\gamma_{\text{container}}$ — the volume coefficient of the vessel ($= 3\alpha$ of the vessel's material); $\gamma_{\text{apparent}}$ — the expansion you actually observe as overflow or a rising column. Read it as bookkeeping: the liquid tries to expand by $\gamma_{\text{real}}$, the container "absorbs" $\gamma_{\text{container}}$ of that by growing with it, and only the remainder shows. For mercury in glass, $\gamma_{\text{real}} = \gamma_{\text{Hg}} = 1.8\times10^{-4}\ \mathrm{K^{-1}}$ and $\gamma_{\text{container}} = 3\alpha_{\text{glass}} = 3\times9\times10^{-6} = 2.7\times10^{-5}\ \mathrm{K^{-1}}$, so the apparent coefficient is $1.8\times10^{-4} - 2.7\times10^{-5} = 1.53\times10^{-4}\ \mathrm{K^{-1}}$.

Heating adds no atoms — mass is conserved — but it inflates the volume, so the same mass spread over more space must become less dense. Starting from density $\rho_0 = m/V_0$ and the expanded volume $V = V_0(1+\gamma\,\Delta T)$:

where: $\rho_0$ — density at the lower temperature; $\gamma$ — the volume coefficient of the substance; the approximation uses $1/(1+x) \approx 1-x$ for tiny $x$. Read it as the mirror image of expansion: whatever fractional volume a body gains, it loses the same fraction in density. Warm mercury by $\Delta T = 100\ \mathrm{K}$ and its density drops by a fraction $\gamma_{\text{Hg}}\,\Delta T = 1.8\times10^{-4}\times100 = 0.018$ — about $1.8\%$ lighter per unit volume. This falling-density-with-warmth is what drives convection: heated fluid becomes less dense, floats up, and carries energy with it — the reason a radiator warms a whole room and a pan of water boils by churning. It is also the exact rule that water famously breaks between 0 and 4 °C, the subject of the companion article on the anomalous expansion of water.

Board level — a steel bridge in the sun (CBSE/ICSE pattern)

A continuous steel girder is $200\ \mathrm{m}$ long at a winter dawn temperature of $10\ \mathrm{^\circ C}$. By a summer afternoon it reaches $50\ \mathrm{^\circ C}$. How much does it lengthen? Take $\alpha_{\text{steel}} = 1.2\times10^{-5}\ \mathrm{K^{-1}}$.

Step 1. Identify the change: $\Delta T = 50 - 10 = 40\ \mathrm{^\circ C} = 40\ \mathrm{K}$ (a difference is the same number on both scales — no conversion needed).

Step 2. Apply equation (1): $\Delta L = L_0\,\alpha\,\Delta T = 200 \times 1.2\times10^{-5} \times 40$.

Step 3. Multiply carefully: $200 \times 1.2\times10^{-5} = 2.4\times10^{-3}$, then $\times 40 = 9.6\times10^{-2}\ \mathrm{m}$. So $\boxed{\Delta L = 0.096\ \mathrm{m} = 9.6\ \mathrm{cm}}$.

Step 4. Sense-check with the fractional form: $\Delta L/L_0 = \alpha\,\Delta T = 4.8\times10^{-4}$, i.e. the girder grows by about $0.048\%$. Small as a fraction, but nearly a $10\ \mathrm{cm}$ gap in absolute terms — which is precisely why bridges ride on sliding expansion joints and long rails are laid with gaps. A rigid clamp instead would build up enormous stress (the subject of the thermal-stress archetype).

Board level — a heated brass sphere (ICSE/IGCSE pattern)

A solid brass sphere has volume $100\ \mathrm{cm^3}$ at room temperature. Its temperature is raised by $50\ \mathrm{K}$. Find the increase in volume. Take $\alpha_{\text{brass}} = 1.9\times10^{-5}\ \mathrm{K^{-1}}$.

Step 1. A sphere is a solid, so use the volume coefficient. First get it from the linear one: $\gamma_{\text{brass}} = 3\alpha_{\text{brass}} = 3 \times 1.9\times10^{-5} = 5.7\times10^{-5}\ \mathrm{K^{-1}}$.

Step 2. Apply equation (3): $\Delta V = V_0\,\gamma\,\Delta T = 100 \times 5.7\times10^{-5} \times 50$.

Step 3. Compute: $100 \times 5.7\times10^{-5} = 5.7\times10^{-3}$, then $\times 50 = 0.285\ \mathrm{cm^3}$. So $\boxed{\Delta V = 0.285\ \mathrm{cm^3}}$, and the new volume is $100.285\ \mathrm{cm^3}$.

Step 4. Note we never needed the radius: expansion cares only about the volume and the temperature change, not the shape. The exam trap here is using $2\alpha$ (that is for area) or forgetting to triple $\alpha$ at all.

Competitive level — a pendulum clock loses time in summer (JEE-Main pattern)

A pendulum clock keeps perfect time in winter. Its steel pendulum warms by $\Delta\theta = 20\ \mathrm{K}$ in summer. How much time does the clock lose per day? Take $\alpha_{\text{steel}} = 1.2\times10^{-5}\ \mathrm{K^{-1}}$; there are $86\,400\ \mathrm{s}$ in a day.

Step 1. Spot the physics: a pendulum's period is $T = 2\pi\sqrt{L/g}$, so $T \propto \sqrt{L}$. A longer pendulum ticks slower, so a warm clock runs slow — it loses time.

Step 2. Turn "how much longer" into "how much slower." Since $T \propto L^{1/2}$, the fractional change in period is half the fractional change in length: $\dfrac{\Delta T_{\text{period}}}{T} = \tfrac12\dfrac{\Delta L}{L} = \tfrac12\,\alpha\,\Delta\theta$.

Step 3. Put in numbers: $\tfrac12\,\alpha\,\Delta\theta = \tfrac12 \times 1.2\times10^{-5} \times 20 = 1.2\times10^{-4}$. Each second of clock-time is now stretched by this fraction.

Step 4. Over a full day the accumulated loss is that fraction of $86\,400\ \mathrm{s}$: $86\,400 \times 1.2\times10^{-4} = \boxed{10.4\ \mathrm{s\ slow\ per\ day}}$ — more than a minute every week. To fix it, clockmakers used low-expansion Invar or self-cancelling gridiron pendulums. Timing note: about two minutes. The whole exam trick is the factor of one-half from $T \propto \sqrt{L}$ — miss it and you double the answer to a wrong $20.7\ \mathrm{s}$. Everything after Step 2 is one multiplication.

Competitive level — how much mercury overflows the glass (JEE-Main/NEET pattern)

A glass vessel of internal volume $1000\ \mathrm{cm^3}$ is filled to the brim with mercury at $0\ \mathrm{^\circ C}$ and heated to $100\ \mathrm{^\circ C}$. How much mercury overflows? Take $\gamma_{\text{Hg}} = 1.8\times10^{-4}\ \mathrm{K^{-1}}$, $\alpha_{\text{glass}} = 9\times10^{-6}\ \mathrm{K^{-1}}$.

Step 1. The overflow is the apparent expansion, so first build the apparent coefficient. The glass cavity's volume coefficient is $\gamma_{\text{glass}} = 3\alpha_{\text{glass}} = 3\times9\times10^{-6} = 2.7\times10^{-5}\ \mathrm{K^{-1}}$.

Step 2. Apply equation (4): $\gamma_{\text{apparent}} = \gamma_{\text{Hg}} - \gamma_{\text{glass}} = 1.8\times10^{-4} - 2.7\times10^{-5} = 1.53\times10^{-4}\ \mathrm{K^{-1}}$.

Step 3. Overflow volume $= V_0\,\gamma_{\text{apparent}}\,\Delta T = 1000 \times 1.53\times10^{-4} \times 100 = \boxed{15.3\ \mathrm{cm^3}}$.

Step 4. Cross-check by bookkeeping: mercury's true expansion is $1000 \times 1.8\times10^{-4} \times 100 = 18\ \mathrm{cm^3}$; the glass cavity grows by $1000 \times 2.7\times10^{-5} \times 100 = 2.7\ \mathrm{cm^3}$; the difference $18 - 2.7 = 15.3\ \mathrm{cm^3}$ overflows — exactly matching Step 3, and the visible $15.3$ is $85\%$ of the true $18$. Timing note: about ninety seconds. The one decision is realising "overflow" means apparent, not real, expansion — students who use $\gamma_{\text{Hg}}$ alone get $18\ \mathrm{cm^3}$ and lose the mark.

The $2\alpha$ / $3\alpha$ swap. Area is $\beta = 2\alpha$, volume is $\gamma = 3\alpha$. The number counts dimensions. Rebuild it by expanding $(1+\alpha\Delta T)^n$ if you ever doubt — it is the most-tested single fact here, appearing in CBSE, ICSE and A-Level derivations every year.

Apparent versus real liquid expansion. "How much overflows" and "what the column shows" are apparent expansion, $\gamma_{\text{real}} - \gamma_{\text{container}}$. Using the liquid's real coefficient alone is the classic mercury-thermometer slip in JEE and NEET.

Thinking a hole shrinks. A heated ring, washer or pipe bore gets larger, because the metal expands outward uniformly. IGCSE and NEET set ring-and-ball questions precisely to catch the opposite guess.

Forgetting density falls. Warm a solid or liquid and its density decreases — volume up, mass fixed. Assertion-reason questions bait "density increases on heating"; it is the convection-driving fact, always a fall.

Confusing $\Delta T$ in °C and K. A temperature difference is numerically identical in kelvin and celsius, so $\Delta T = 40\ \mathrm{^\circ C} = 40\ \mathrm{K}$ — never add $273$ to a change. You only convert absolute temperatures, never differences.

Dropping the factor of one-half in clock problems. Because period $\propto \sqrt{L}$, the fractional timing error is $\tfrac12\alpha\Delta\theta$, not $\alpha\Delta\theta$. Halve it — this is the number-one error in the pendulum-clock numerical.

Which value of the coefficient? Boards quote $\alpha$ in $\mathrm{K^{-1}}$ (or the identical $\mathrm{^\circ C^{-1}}$); some tables give $\gamma$ directly for liquids. Use exactly the value the paper supplies, and check whether you are handed the linear ($\alpha$) or the volume ($\gamma$) coefficient before you triple anything.

Texts. NCERT Physics Class 11, Ch. 11 (Thermal Properties of Matter) — the CBSE/JEE statement of $\alpha$, $\beta$, $\gamma$ and apparent expansion · Cambridge IGCSE Physics 0625, Thermal Physics (expansion of solids, liquids and gases) · ISC/ICSE Concise Physics, Heat chapter · Halliday, Resnick & Walker, "Temperature, Heat and the First Law" for the potential-well picture.

Provenance. Standard results. A verifier should recompute: $\beta_{\text{steel}} = 2\times1.2\times10^{-5} = 2.4\times10^{-5}$ and the plate gain $1\times2.4\times10^{-5}\times40 = 9.6\times10^{-4}\ \mathrm{m^2}$; $\gamma_{\text{brass}} = 3\times1.9\times10^{-5} = 5.7\times10^{-5}$; the binomial check $(1+1.2\times10^{-3})^2$ and $(1+1.2\times10^{-3})^3$ against $1+2(1.2\times10^{-3})$ and $1+3(1.2\times10^{-3})$, ratio $6\times10^{-4}$; bridge $\Delta L = 200\times1.2\times10^{-5}\times40 = 0.096\ \mathrm{m}$, fraction $4.8\times10^{-4}$; sphere $\Delta V = 100\times5.7\times10^{-5}\times50 = 0.285\ \mathrm{cm^3}$; apparent $\gamma_{\text{app}} = 1.8\times10^{-4}-2.7\times10^{-5} = 1.53\times10^{-4}$, visible fraction $0.85$, overflow $1000\times1.53\times10^{-4}\times100 = 15.3\ \mathrm{cm^3}$ (true $18$, glass $2.7$); density drop $1.8\times10^{-4}\times100 = 0.018$; pendulum $\tfrac12\times1.2\times10^{-5}\times20 = 1.2\times10^{-4}$ and $86\,400\times1.2\times10^{-4} = 10.4\ \mathrm{s/day}$; Invar ratio $\alpha_{\text{steel}}/\alpha_{\text{Invar}} = 10$.

Prerequisites. Thermometry & Temperature Scales (what $\Delta T$ means and why K = °C for a difference). Leads to. Anomalous Expansion of Water (the great exception) · Thermal Stress & Bimetallic Strips (what happens when expansion is prevented). Go deeper (UG). The temperature-dependence of the coefficient and the full anharmonic-oscillator treatment of the interatomic potential belong to the university solid-state and statistical-mechanics tracks.