Thermal expansion

Leave a gap at the joints of a railway track, or a metal lid stuck on a glass jar loosens under hot water — both are everyday sightings of thermal expansion, the tendency of matter to grow in length, area, or volume as its temperature rises. Heating a substance does not add new material to it; it makes the particles jostle more violently and settle, on average, a little farther apart, so the whole body swells. Solids expand along a length, over an area, and through a volume; liquids and gases, having no fixed shape, only expand in volume. This article builds the three coefficients of expansion, the neat algebraic relation between them for solids, the real-versus-apparent expansion of a liquid in a container, and the curious case of water, which shrinks rather than grows between 0°C and 4°C.

The core insight

Heating increases interatomic spacing, so length, area, and volume all grow in direct proportion to the temperature rise, via the coefficients α\alpha, β=2α\beta = 2\alpha, and γ=3α\gamma = 3\alpha for isotropic solids.

The core idea

Why expansion happens. Atoms in a solid sit near the bottom of a potential-energy well and vibrate about a fixed mean position. That well is not perfectly symmetric — it is easier to push atoms apart than to squeeze them together — so as temperature rises and vibration amplitude grows, the average separation increases, not just the wiggle. Stack that tiny per-atom growth over billions of atoms in a row and a visible, measurable expansion results.

Linear expansion. For a rod of original length L0L_0 heated through a temperature change ΔT\Delta T, experiment shows the change in length is directly proportional to both the original length and the temperature change:

ΔL=αL0ΔT\Delta L = \alpha L_0 \Delta T

where: ΔL\Delta L is the increase in length, L0L_0 is the original length, ΔT\Delta T is the temperature rise, and α\alpha (unit K−1\text{K}^{-1} or ∘C−1^\circ\text{C}^{-1}) is the coefficient of linear expansion, a fixed property of the material, typically of order 10−5 ∘C−110^{-5}\ ^\circ\text{C}^{-1} for common metals. The new length is L=L0(1+αΔT)L = L_0(1 + \alpha \Delta T).

Area and volume. A flat sheet of the same isotropic solid expands in every direction at once, so its area grows according to a coefficient of superficial (areal) expansion β\beta, with ΔA=βA0ΔT\Delta A = \beta A_0 \Delta T; a solid block expands in three directions, governed by the coefficient of cubical (volume) expansion γ\gamma, with ΔV=γV0ΔT\Delta V = \gamma V_0 \Delta T. Because area and volume are just length multiplied by itself two or three times, β\beta and γ\gamma are not independent new constants for an isotropic solid — they are fixed multiples of α\alpha, derived below.

Liquids, gases, and apparent expansion. Liquids and gases have no shape of their own, so only γ\gamma is meaningful for them, and their γ\gamma values are usually far larger than a solid's, since intermolecular bonds are weaker and give more easily. A liquid stored in a solid container shows a smaller apparent expansion than its true (real, or absolute) expansion, because the container is expanding too and quietly making room:

γreal=γapparent+γcontainer\gamma_{\text{real}} = \gamma_{\text{apparent}} + \gamma_{\text{container}}
Figure
A rod at a lower temperature shown solid and the same rod at a higher temperature shown longer with a dashed extension marked delta L, next to a small square that has grown into a slightly larger square to show area expansion in two directions at once
One-dimensional versus two-dimensional expansion rod: length grows by alpha L0 delta T L0 at T L at T plus delta T delta L sheet: area grows by beta A0 delta T grows outward on every side Same physical stretching, counted along one axis, two axes, or three

Anomalous water. Almost every substance expands smoothly on heating, but water between 0°C and 4°C does the opposite: it contracts, reaching maximum density at 4°C, and only expands normally above that. This is why a pond freezes from the top down — the coldest, lightest water (near 0°C) stays on the surface while the densest water (at 4°C) sinks and insulates life below.

Derivation AWhy area and volume expand at 2 alpha and 3 alpha

The goal is to show, using nothing more than the linear expansion law and elementary algebra (dropping products of small quantities), that for an isotropic solid the areal coefficient must equal exactly twice the linear coefficient and the volumetric coefficient exactly three times it — so a single measured α\alpha is enough to predict β\beta and γ\gamma without any separate experiment.

Step 1. Start from a square sheet of side L0L_0, so its original area is A0=L02A_0 = L_0^2. Because the sheet is isotropic, heating through ΔT\Delta T stretches every linear dimension by the same factor, so each side becomes L=L0(1+αΔT)L = L_0(1 + \alpha \Delta T), exactly as for the rod.

Step 2. Square this new side to get the new area. A=L2=L02(1+αΔT)2=L02(1+2αΔT+α2ΔT2)A = L^2 = L_0^2(1+\alpha\Delta T)^2 = L_0^2\left(1 + 2\alpha\Delta T + \alpha^2\Delta T^2\right). Since α\alpha is of order 10−5 ∘C−110^{-5}\ ^\circ\text{C}^{-1}, the term α2ΔT2\alpha^2\Delta T^2 is smaller than the linear term by another factor of αΔT\alpha \Delta T — utterly negligible even for a 1000∘C1000^\circ\text{C} rise — so it is dropped.

Step 3. Read off the change. A≈L02(1+2αΔT)=A0+2αA0ΔTA \approx L_0^2(1 + 2\alpha\Delta T) = A_0 + 2\alpha A_0 \Delta T, so ΔA=2αA0ΔT\Delta A = 2\alpha A_0 \Delta T. Comparing with the defining relation ΔA=βA0ΔT\Delta A = \beta A_0 \Delta T gives β=2α\beta = 2\alpha directly — the factor of 22 is simply the two independent directions in a plane, each contributing one power of αΔT\alpha \Delta T.

Step 4. Repeat for a cube of side L0L_0 and volume V0=L03V_0 = L_0^3. Cubing L=L0(1+αΔT)L = L_0(1+\alpha\Delta T) and again discarding every term with α2\alpha^2 or α3\alpha^3 in it leaves V≈L03(1+3αΔT)V \approx L_0^3(1 + 3\alpha\Delta T), so ΔV=3αV0ΔT\Delta V = 3\alpha V_0 \Delta T, and comparing with ΔV=γV0ΔT\Delta V = \gamma V_0 \Delta T gives γ=3α\gamma = 3\alpha — the factor of 33 counts the three independent directions of space. These relations hold only for isotropic solids that expand equally in all directions; they do not apply to liquids or gases, which have no meaningful α\alpha or β\beta at all.

Worked example

A steel rod has a length of 2.000 m at 20°C, and steel has α=1.2×10−5 ∘C−1\alpha = 1.2\times10^{-5}\ ^\circ\text{C}^{-1}. Find its length when heated to 120°C, and the volume expansion of a steel cube of the same side length under the same heating.

Step 1. Identify the data: L0=2.000L_0 = 2.000 m, ΔT=120−20=100∘C\Delta T = 120 - 20 = 100^\circ\text{C}, α=1.2×10−5 ∘C−1\alpha = 1.2\times10^{-5}\ ^\circ\text{C}^{-1}.

Step 2. Apply the linear expansion law: ΔL=αL0ΔT=(1.2×10−5)(2.000)(100)=2.4×10−3 m=2.4 mm\Delta L = \alpha L_0 \Delta T = (1.2\times10^{-5})(2.000)(100) = 2.4\times10^{-3}\ \text{m} = 2.4\ \text{mm}, giving L=2.000+0.0024=2.0024 mL = 2.000 + 0.0024 = 2.0024\ \text{m}. This is why long steel bridges and railway tracks are built with expansion gaps: a full-length girder can shift by several centimetres over such a temperature swing.

Competitive stretch

Step 3. Now take a steel cube of side 2.000 m (volume V0=2.0003=8.000 m3V_0 = 2.000^3 = 8.000\ \text{m}^3) heated through the same ΔT=100∘C\Delta T = 100^\circ\text{C}. First find γ=3α=3(1.2×10−5)=3.6×10−5 ∘C−1\gamma = 3\alpha = 3(1.2\times10^{-5}) = 3.6\times10^{-5}\ ^\circ\text{C}^{-1}.

Step 4. Apply the volume expansion law: ΔV=γV0ΔT=(3.6×10−5)(8.000)(100)=0.0288 m3\Delta V = \gamma V_0 \Delta T = (3.6\times10^{-5})(8.000)(100) = 0.0288\ \text{m}^3, about 28.828.8 litres. The cube's volume grows by 0.0288 m30.0288\ \text{m}^3, nine times the fractional growth you might naively expect from tripling only the linear answer, because volume compounds the stretch in three directions at once.

Common traps

  • Applying β=2α\beta = 2\alpha and γ=3α\gamma = 3\alpha to liquids or gases — these relations hold only for isotropic solids, since liquids and gases have no fixed shape and no meaningful α\alpha or β\beta.
  • Mixing up apparent expansion of a liquid (what you observe, smaller) with its real expansion (what the liquid alone would do, larger) — the container's own expansion always eats into the observed rise.
  • Assuming water always expands on heating — from 0°C to 4°C it contracts and reaches maximum density at 4°C; this reversal is a favourite conceptual question.
  • Needlessly converting ΔT\Delta T between °C and kelvin — since both scales have equal-sized degrees, ΔT\Delta T has the same numeric value in either, so converting only invites sign slips.

Ask the tutor: if a numerical on real-versus-apparent liquid expansion or the alpha-beta-gamma relations feels shaky, ask the AI tutor to work through one more example with you step by step.

Why is gamma exactly three times alpha and not some other number depending on the shape of the object?

The factor of three comes purely from counting dimensions, not from the object's shape. Any solid, however irregularly shaped, can be thought of as built from tiny cubes; each cube's volume grows by the (1+αΔT)3≈1+3αΔT(1+\alpha\Delta T)^3 \approx 1+3\alpha\Delta T factor derived from stretching every edge equally. Summing over all the tiny cubes still gives γ=3α\gamma = 3\alpha for the whole object, as long as the material is isotropic — meaning it expands equally in every direction, which is true for most everyday solids like steel or glass, though not for some crystals.

If the container also expands when a liquid is heated, why does the liquid still overflow at all?

Because liquids almost always have a far larger γ\gamma than the solid container holding them — mercury's coefficient is roughly six to seven times that of glass. Even though the glass genuinely does expand and creates some extra room, the mercury's own volume grows so much faster that there is not enough new room to hold it all, and the excess spills over. The apparent expansion you see is the small leftover after subtracting the container's contribution from the liquid's much larger real expansion.

Why does ice float, and how is that connected to water's anomalous expansion between 0 and 4 degrees Celsius?

Water reaches its maximum density at 4°C; below that, right down to 0°C, it actually expands again as it cools further, and ice is even less dense still, which is why it floats. In a cooling pond, water near 4°C is densest and sinks to the bottom, while colder water and eventually ice stay near the surface. That surface layer of ice then insulates the water beneath it, keeping it liquid near 4°C through winter — which is exactly why fish and other aquatic life can survive under a frozen pond.

Before this

Where this leads