Thermal expansion

Look closely at a railway line on a hot afternoon and you will see a small gap left between one steel segment and the next — not a manufacturing flaw but a deliberate allowance for a fact of nature: almost every solid, liquid and gas expands when it is heated. Bridges roll on sliding bearings, concrete highways are poured in slabs with tar-filled joints, and the loose wire of an overhead power line sags a little more in summer than in winter, all because of the same physics. Even the humble mercury thermometer works only because mercury's steady volume expansion with temperature gives it a column that climbs in exact proportion to how hot it is. The underlying picture is simple and mechanical: heating a solid makes its atoms jiggle more violently about their fixed positions in the lattice, and because the interatomic force curve is not perfectly symmetric, the average separation between neighbouring atoms grows as the vibration amplitude grows. Millions of such tiny increases, added along the length of a rod, across the face of a plate, or through the bulk of a block, show up as measurable linear, areal and volume expansion — three faces of the same microscopic effect, related to each other by nothing more than counting dimensions. This article builds that picture into three clean formulas — $\Delta L = \alpha L\,\Delta T$ , $\Delta A = \beta A\,\Delta T$ , $\Delta V = \gamma V\,\Delta T$ — derives the working relations $\beta \approx 2\alpha$ and $\gamma \approx 3\alpha$ from first principles, applies them to a real railway rail and a heated brass sphere, explains why a hole in a heated plate gets bigger rather than smaller, and closes with the curious case of water, which contracts instead of expanding as it is warmed from $0^\circ\text{C}$ to $4^\circ\text{C}$ — a quirk of nature that keeps fish alive under frozen lakes.

Key takeaways

Linear expansion. A rod of length $L$ heated through $\Delta T$ grows by $\Delta L = \alpha L\,\Delta T$ , where $\alpha$ is the coefficient of linear expansion, unit $\text{K}^{-1}$ .
Areal and volume expansion follow the same pattern. $\Delta A = \beta A\,\Delta T$ and $\Delta V = \gamma V\,\Delta T$ , with $\beta$ and $\gamma$ the coefficients of superficial and cubical expansion.
The three coefficients are linked. For an isotropic solid, $\beta \approx 2\alpha$ and $\gamma \approx 3\alpha$ — because area scales with two lengths and volume with three.
Liquids only have a volume coefficient, since they have no fixed shape; heating a liquid lowers its density, $\rho' \approx \rho(1-\gamma\,\Delta T)$ .
A cavity inside a solid expands exactly like a solid piece of the same material would — heating a metal ring with a hole makes the hole bigger, not smaller.
Water is the great exception. Between $0^\circ\text{C}$ and $4^\circ\text{C}$ it contracts on heating, reaching maximum density at $4^\circ\text{C}$ — this anomalous expansion is why ice floats and lakes freeze from the top.
Key idea: heating a solid, liquid, or gas increases the average spacing between its particles, so length, area, and volume all grow linearly with temperature change through $\Delta L=\alpha L\Delta T$ , $\Delta A=\beta A\Delta T=2\alpha A\Delta T$ , and $\Delta V=\gamma V\Delta T=3\alpha V\Delta T$ .

The core insight

Heating a solid increases the average interatomic spacing, so its length, area, and volume each grow in proportion to the original size and the temperature rise, giving $\Delta L=\alpha L\,\Delta T$ , $\Delta A\approx2\alpha A\,\Delta T$ , and $\Delta V\approx3\alpha V\,\Delta T$ , with water's anomalous contraction between $0^\circ\text{C}$ and $4^\circ\text{C}$ as the one notable exception.

Where it fits in the syllabus

Thermal expansion is one of the four prescribed strands of the Thermal Properties of Matter chapter in CBSE Class 11 Physics, alongside thermometry, specific heat/calorimetry, and heat transfer. The board syllabus asks for the three expansion coefficients $\alpha$ , $\beta$ , $\gamma$ , the approximate relations connecting them, the expansion of liquids and its effect on density, and the anomalous expansion of water — every one of those is covered here in full board depth. Newton's law of cooling and greenhouse-effect discussions belong to the heat-transfer strand of the same chapter, not here, and are only mentioned where a signpost is useful.

The physical picture

At the atomic scale, a solid is a lattice of atoms held near fixed equilibrium positions by interatomic forces that behave, for small displacements, almost like springs. When the solid is heated, each atom vibrates with larger amplitude about its equilibrium site. Crucially, the potential-energy curve between neighbouring atoms is not a perfect symmetric parabola — it rises more steeply on the "too close" side than it falls on the "too far" side, because squeezing atoms together meets strong repulsion while pulling them apart meets weaker attraction. The result is that the average position of a vibrating atom shifts slightly outward as the vibration amplitude grows, so the mean interatomic distance increases with temperature. Summed along the length of a rod, this microscopic asymmetry produces a macroscopic, measurable elongation. For a solid rod of original length $L$ at temperature $T$ , whose temperature is raised by $\Delta T$ , experiment shows the fractional change in length is directly proportional to $\Delta T$ for the modest temperature ranges met in the classroom:

\Delta L = \alpha\, L\, \Delta T \qquad\text{so that}\qquad L' = L\left(1+\alpha\,\Delta T\right)

where: $\Delta L$ is the increase in length, $L$ is the original length, $\Delta T$ is the temperature rise (in kelvin or degree Celsius — a difference is the same in both scales), and $\alpha$ , the coefficient of linear expansion, is a material property with SI unit $\text{K}^{-1}$ (equivalently $^\circ\text{C}^{-1}$ ). Physically, $\alpha$ tells you the fractional stretch per unit temperature rise; steel's $\alpha \approx 1.2\times10^{-5}\ \text{K}^{-1}$ means a one-metre steel bar lengthens by only about twelve micrometres for every one-kelvin rise — a tiny effect per degree that becomes centimetres over tens of metres and tens of degrees, exactly the scale railway engineers must plan for. A flat solid also expands in area, and a bulk solid expands in volume, each following the identical linear-in- $\Delta T$ pattern:

\Delta A = \beta\, A\, \Delta T \qquad\qquad \Delta V = \gamma\, V\, \Delta T

where: $\beta$ is the coefficient of areal (superficial) expansion and $\gamma$ is the coefficient of volume (cubical) expansion, both also in $\text{K}^{-1}$ . For an isotropic solid — one whose atoms are, on average, arranged the same way in every direction, so it expands equally along every axis — these three coefficients are not independent; they are tied together by simple dimension-counting, giving the working relations

\beta \approx 2\alpha \qquad\qquad \gamma \approx 3\alpha

Liquids, having no fixed shape, possess only a volume coefficient $\gamma$ ; talking about a liquid's "linear expansion" is meaningless because a liquid has no fixed length to begin with. Since density is $\rho = m/V$ and mass does not change on heating, an increase in volume must lower density:

\rho' = \frac{\rho}{1+\gamma\,\Delta T} \approx \rho\left(1-\gamma\,\Delta T\right) \quad\text{for small } \gamma\,\Delta T

where: $\rho$ is the original density and $\rho'$ the density after heating; the approximation on the right drops terms of order $(\gamma\,\Delta T)^2$ and higher, which are negligible because $\gamma\,\Delta T$ is typically a small number (of order $10^{-2}$ or smaller) for the temperature ranges tested at this level.

Figure

A rod at a cooler temperature and the same rod slightly longer at a hotter temperature, with the increase in length shown as delta L equals alpha L delta T

Thermal expansion is not just a textbook curiosity — it drives real devices. A bimetallic strip, made of two metals with different values of $\alpha$ (commonly brass and steel) riveted or welded together along their length, bends when heated because the metal with the larger $\alpha$ stretches more than the metal with the smaller $\alpha$ , forcing the strip to curl toward the low-expansion side. This bending is harnessed in thermostats and fire alarms, where the curling strip makes or breaks an electrical contact automatically at a preset temperature — thermal expansion converted directly into mechanical motion and, from there, into a switch. A subtlety worth internalising early: a hole or cavity in a solid expands exactly as if the hole were filled with the same solid material. If a flat metal plate with a circular hole is heated uniformly, every linear dimension of the plate — including the diameter of the hole — grows by the factor $(1+\alpha\,\Delta T)$ . It is tempting to imagine the surrounding metal "closing in" on the hole, but the correct picture is that the whole pattern, hole included, expands proportionally, so the hole gets larger, never smaller, on heating a plate of positive $\alpha$ .

Derivation AGetting beta approximately equal to 2 alpha and gamma approximately equal to 3 alpha from linear expansion

The goal is to show, using nothing beyond the binomial approximation and simple geometry, that the areal and volume expansion coefficients of an isotropic solid are not independent constants to be memorised separately but are fixed once $\alpha$ is known, because area scales as the square of a length and volume scales as the cube of a length; the derivation matters because it lets a student reconstruct $\beta$ and $\gamma$ on the spot instead of recalling three numbers.

Step 1. Set up a square plate of side $L$ in an isotropic solid, so its area is $A = L^2$ . When the solid is heated through $\Delta T$ , every linear dimension grows by the same factor, because isotropy means the material expands identically along any direction: the new side is $L' = L(1+\alpha\,\Delta T)$ .

Step 2. Square the new side to get the new area. $A' = (L')^2 = L^2(1+\alpha\,\Delta T)^2 = L^2\left(1 + 2\alpha\,\Delta T + \alpha^2\Delta T^2\right)$ . Since $\alpha$ is of order $10^{-5}\ \text{K}^{-1}$ and $\Delta T$ is at most a few hundred kelvin in any realistic problem, the product $\alpha\,\Delta T$ is of order $10^{-3}$ or smaller, so the term $\alpha^2\Delta T^2$ is smaller still, of order $10^{-6}$ , and can be dropped without measurable error.

Step 3. Drop the negligible squared term. $A' \approx L^2\left(1+2\alpha\,\Delta T\right) = A\left(1+2\alpha\,\Delta T\right)$ . Comparing with the defining relation $A' = A(1+\beta\,\Delta T)$ term by term gives $\beta = 2\alpha$ directly — the areal coefficient is twice the linear one because area carries two powers of length, and each power contributes one factor of $\alpha\,\Delta T$ to first order.

Step 4. Repeat for a cube of side $L$ , so $V = L^3$ . The new volume is $V' = (L')^3 = L^3(1+\alpha\,\Delta T)^3 = L^3\left(1+3\alpha\,\Delta T+3\alpha^2\Delta T^2+\alpha^3\Delta T^3\right) \approx L^3\left(1+3\alpha\,\Delta T\right) = V(1+3\alpha\,\Delta T)$ , where again all terms beyond the first power of $\alpha\,\Delta T$ are negligible. Comparing with $V' = V(1+\gamma\,\Delta T)$ gives $\gamma = 3\alpha$ , because volume carries three powers of length, each contributing one factor of $\alpha\,\Delta T$ .

Step 5. Read the pattern and its limits. The relations $\beta\approx2\alpha$ and $\gamma\approx3\alpha$ are approximations, exact only in the limit of infinitesimally small $\alpha\,\Delta T$ , and they assume the solid is isotropic (a crystal that expands differently along different axes needs separate coefficients along each axis, which is beyond this syllabus). For the temperature changes typical of board and competitive problems, the approximation is accurate to a fraction of a percent, which is why it is used freely in numerical work.

Worked examples

Board level - railway rail expansion

A steel rail is $30.0\ \text{m}$ long when laid at a temperature of $10^\circ\text{C}$ . The coefficient of linear expansion of steel is $\alpha = 1.2\times10^{-5}\ \text{K}^{-1}$ . Find the increase in length when the temperature rises to $45^\circ\text{C}$ on a hot day.

Step 1. Find the temperature change. $\Delta T = 45^\circ\text{C} - 10^\circ\text{C} = 35\ \text{K}$ — a Celsius difference converts directly to a kelvin difference, since both scales have identically sized degrees.

Step 2. Apply the linear expansion formula. $\Delta L = \alpha L\,\Delta T = (1.2\times10^{-5}\ \text{K}^{-1})(30.0\ \text{m})(35\ \text{K})$ .

Step 3. Multiply through. $\Delta L = (1.2\times10^{-5})(1050\ \text{m}\,\cdot\,\text{K}) = 1.26\times10^{-2}\ \text{m}$ .

The rail lengthens by $\boxed{1.26\times10^{-2}\ \text{m} = 1.26\ \text{cm}}$ — this is precisely the order of gap engineers must leave between successive rail segments so the expanding rail does not buckle the track.

Competitive level - brass sphere volume and a hidden hole check

A brass sphere has a volume of $500\ \text{cm}^3$ at $20^\circ\text{C}$ , with $\alpha_{\text{brass}} = 1.9\times10^{-5}\ \text{K}^{-1}$ . Find its volume at $120^\circ\text{C}$ , and also find the new diameter of a circular hole of diameter $2.00\ \text{cm}$ drilled through a brass plate at the same starting temperature, heated over the same range.

Step 1. Convert the linear coefficient to the volume coefficient using the isotropic relation derived above. $\gamma \approx 3\alpha = 3\times(1.9\times10^{-5}) = 5.7\times10^{-5}\ \text{K}^{-1}$ .

Step 2. Find the temperature change and the volume increase. $\Delta T = 120-20 = 100\ \text{K}$ , so $\Delta V = \gamma V\,\Delta T = (5.7\times10^{-5})(500\ \text{cm}^3)(100\ \text{K}) = 2.85\ \text{cm}^3$ .

Step 3. Add to the original volume. $V' = 500 + 2.85 = 502.85\ \text{cm}^3$ .

Step 4. For the hole, apply linear expansion directly to the diameter, treating the hole as if it were solid brass of the same material — the key idea being that a cavity expands exactly like the material that would fill it. $\Delta d = \alpha d\,\Delta T = (1.9\times10^{-5})(2.00\ \text{cm})(100\ \text{K}) = 3.8\times10^{-3}\ \text{cm}$ , so the hole's new diameter is $2.00 + 3.8\times10^{-3} = 2.0038\ \text{cm}$ — bigger, not smaller.

Sphere volume: $\boxed{502.85\ \text{cm}^3}$ ; hole diameter: $\boxed{2.0038\ \text{cm}}$ (an increase, confirming the hole enlarges on heating). Timing note: on a competitive-exam clock, do not re-derive $\gamma=3\alpha$ from scratch — quote it, plug numbers, and move on; a full derivation is only expected if the question explicitly asks you to prove the relation. Aim for under ninety seconds on a plug-in question of this type.

Common traps

Treating $\beta=2\alpha$ and $\gamma=3\alpha$ as exact laws rather than first-order approximations — they hold only because $\alpha\,\Delta T$ is small in practical problems; a question with an unusually large $\Delta T$ is signalling that you should keep the derivation's higher-order terms in mind, even though they are almost never numerically demanded at this level.
Assuming a hole in a heated solid shrinks — it enlarges, following the same expansion law as the surrounding material, because the missing material would itself have expanded had it been there.
Forgetting that liquids have only a volume coefficient $\gamma$ — questions asking for the "linear expansion of water" are testing whether you notice a liquid has no fixed length or shape to expand linearly.
Assuming every liquid, including water, always expands on heating — water between $0^\circ\text{C}$ and $4^\circ\text{C}$ is the standard exception, and exam-setters love to test exactly this range.
Mixing up $\Delta T$ in kelvin with an absolute temperature — a temperature difference is numerically identical in Celsius and kelvin, so never add $273$ to a $\Delta T$ ; adding $273$ is only ever done to convert an absolute Celsius temperature to an absolute kelvin temperature, never to a difference.

Common trapWater is not "just another liquid" between 0 and 4 degrees Celsius

Most liquids expand steadily as they are heated, and it is tempting to assume water does too across its whole liquid range. In fact, between $0^\circ\text{C}$ and $4^\circ\text{C}$ , heating water makes it contract: its volume shrinks and its density rises, reaching a maximum density at $4^\circ\text{C}$ . Only above $4^\circ\text{C}$ does water resume the normal pattern of expanding as it is heated further. This anomalous behaviour means that as a lake cools toward winter, the coldest surface water (below $4^\circ\text{C}$ ) stays less dense than the $4^\circ\text{C}$ water beneath it and does not sink, so ice forms at the top while liquid water at about $4^\circ\text{C}$ remains at the bottom — insulating aquatic life from freezing solid. A question that asks about water "cooling from $10^\circ\text{C}$ to $2^\circ\text{C}$ " is deliberately crossing the anomalous region, and a straight application of $\Delta V=\gamma V\,\Delta T$ with a single constant $\gamma$ will give the wrong sign of volume change for the part of the journey below $4^\circ\text{C}$ .

You should be able to…

Write and correctly use $\Delta L=\alpha L\,\Delta T$ , $\Delta A=\beta A\,\Delta T$ , and $\Delta V=\gamma V\,\Delta T$ , identifying each coefficient's unit.
Derive $\beta\approx2\alpha$ and $\gamma\approx3\alpha$ from the binomial approximation for an isotropic solid.
Calculate the change in length, area, or volume of a solid, or the change in a liquid's density, for a given temperature change.
Explain, with a reason, why a hole or cavity in a heated solid enlarges rather than shrinks.
Describe the anomalous expansion of water between $0^\circ\text{C}$ and $4^\circ\text{C}$ and connect it to why lakes freeze from the surface down.
Recognise when a bimetallic-strip question is testing thermal expansion and explain the direction it bends.

Ask the tutor: want a guided derivation of why a hole enlarges rather than shrinks, or more practice converting between $\alpha$ , $\beta$ , and $\gamma$ across mixed solid-and-liquid problems — ask the AI tutor for a walkthrough.

Why does alpha have units of per kelvin instead of just being a pure number?

Because $\alpha$ is defined as a fractional change in length per unit temperature rise: $\alpha = \dfrac{\Delta L/L}{\Delta T}$ . The numerator, $\Delta L/L$ , is a pure ratio (metres over metres), so it has no units of its own. Dividing that dimensionless ratio by a temperature change $\Delta T$ (measured in kelvin) leaves units of one-over-kelvin, or $\text{K}^{-1}$ . This is exactly why you can also write $\alpha$ in $^\circ\text{C}^{-1}$ without changing its numerical value — a difference of one kelvin is identical in size to a difference of one degree Celsius, even though the two scales have different zero points.

If beta is approximately 2 alpha, why isn't it exactly 2 alpha?

The exact relation, worked out by fully expanding $(1+\alpha\,\Delta T)^2$ , is $\beta = 2\alpha + \alpha^2\,\Delta T$ . For ordinary values of $\alpha$ (around $10^{-5}\ \text{K}^{-1}$ ) and ordinary temperature changes (tens to a few hundred kelvin), the extra term $\alpha^2\,\Delta T$ is roughly a million times smaller than $2\alpha$ , so it is dropped and $\beta\approx2\alpha$ is an excellent approximation. It would only start to matter for enormous, unrealistic values of $\Delta T$ , which is why CBSE numericals never push you into that regime.

Does thermal expansion happen in gases too, and why isn't it covered the same way here?

Yes — gases expand on heating far more dramatically than solids or liquids, but a gas has no fixed volume of its own at all; its volume depends jointly on pressure and temperature through the ideal gas equation $PV=nRT$ , which you meet in the kinetic theory chapter. The single coefficient $\gamma$ used for solids and liquids is not the natural way to describe a gas, so CBSE treats gas expansion separately through the gas laws rather than through $\alpha$ , $\beta$ , $\gamma$ .

Why do railway tracks and bridges need expansion gaps if the expansion is only a tiny fraction of a percent?

A tiny fractional change becomes a large absolute change over a long enough length. Steel's $\alpha\approx1.2\times10^{-5}\ \text{K}^{-1}$ means each metre stretches by only about twelve micrometres per kelvin, but a rail is metres long and can see a temperature swing of thirty or forty kelvin between a cold night and a hot afternoon — as the worked example shows, that adds up to over a centimetre for a thirty-metre rail. Without a gap, thousands of rail segments each pushing against their neighbours would build up enormous compressive stress, enough to buckle the track sideways, which is exactly the failure mode expansion joints are designed to prevent.