Linear Thermal Expansion of Copper theory

The physical origin of thermal expansion in crystalline solids is rooted in the asymmetric shape of the interatomic potential energy curve U(r). In a realistic crystal, the interatomic potential can be modeled by a Lennard-Jones (12-6) or Morse function: U(r) = U_0 [(r_0/r)¹² − 2(r_0/r)⁶], or expanded as a Taylor series about the equilibrium separation r_0: U(x) = c x² − g x³ − f x⁴, where x = r − r_0 represents atomic displacement, c is the harmonic restoring force constant, g is the cubic anharmonicity coefficient, and f is the quartic coefficient. If the potential were strictly harmonic (g = 0, f = 0), the Boltzmann distribution P(x) ∝ exp[−U(x)/(kT)] would be symmetric in ±x, meaning the expectation value ⟨x⟩ = ∫ x exp[−U(x)/(kT)] dx / ∫ exp[−U(x)/(kT)] dx would remain identically zero at all temperatures. However, because atomic electron shells strongly resist overlapping (producing a steep repulsive wall for x < 0) while attractive bonding forces decay gradually with distance (a shallow trough for x > 0), the cubic anharmonic term g is strictly positive. Evaluating the thermal expectation value using perturbation theory yields: ⟨x⟩ ≈ (3g / 4c²) · kT. Thus, as temperature T rises, the average interatomic separation increases linearly with T, giving rise to macroscopic thermal expansion. Without lattice anharmonicity, solids would not expand when heated.

In isotropic materials where physical properties are identical in all spatial directions, thermal dilatation is described by a single scalar coefficient. Consider a rectangular block of dimensions L_x, L_y, L_z at baseline temperature T_0. When heated through ΔT, each linear dimension expands according to L_i(T) = L_{i,0}(1 + α ΔT). The resulting area of any face (e.g., A = L_x L_y) becomes A(T) = L_{x,0} L_{y,0}(1 + α ΔT)² = A_0(1 + 2α ΔT + α² ΔT²). Because typical expansion coefficients for metals are of order 10⁻⁵ K⁻¹, the second-order term α² ΔT² is negligible (e.g., (1.68 × 10⁻⁵ × 50)² ≈ 7 × 10⁻⁷), allowing the areal thermal expansion coefficient to be approximated as β ≈ 2α. Similarly, the expanded volume evaluates to V(T) = V_0(1 + α ΔT)³ = V_0(1 + 3α ΔT + 3α² ΔT² + α³ ΔT³). Neglecting higher-order terms yields V(T) ≈ V_0(1 + γ ΔT), establishing the fundamental thermodynamic relation γ ≈ 3α for isotropic solids. For anisotropic crystals such as calcite, zinc, or graphite, thermal expansion is a second-rank tensor α_ij, where expansion along different crystallographic axes can vary dramatically and may even exhibit negative thermal expansion along specific axes.

Measuring the thermal expansion of a 1.00-meter rod over a temperature elevation of 50 °C poses a delicate metrological challenge: the total elongation is only ΔL = 1.000 m × 16.80 × 10⁻⁶ K⁻¹ × 50 K = 0.840 mm (840 microns). If a mechanical micrometer screw is turned until it physically pinches the expanding rod, the compressive contact force can easily distort the frame, indent the soft copper rod tip, and compress the specimen longitudinally according to Hooke's law: δ_comp = F L_0 / (A Y). For a 6 mm diameter copper rod (A ≈ 2.83 × 10⁻⁵ m², Young's modulus Y ≈ 110 GPa), an anvil contact force of just 10 newtons causes a compressive strain of δ_comp ≈ 3.2 microns—introducing a 0.4% systematic measurement error. To eliminate contact force entirely, modern university steam dilatometers employ an electrical indicator circuit. The micrometer screw and specimen rod are placed in series with a 1.5 V DC source, a high-value current-limiting resistor, and a sensitive galvanometer or LED. As the micrometer screw is advanced slowly, the very first atomic contact between the polished spindle face and rod end closes the electrical circuit, instantly triggering a galvanometer deflection or lighting the indicator. The observer records the drum reading at this exact threshold of zero mechanical pressure, ensuring sub-micron accuracy.

When a solid bar is heated while its ends are rigidly constrained between unyielding walls, its natural tendency to expand is frustrated, generating immense internal compressive stress. The free thermal strain that would occur without constraint is ε_th = ΔL/L_0 = α ΔT. To maintain the bar at its original length (net strain ε_net = 0), the rigid supports must exert an opposing mechanical compressive strain equal in magnitude to the thermal strain: ε_mech = −σ / Y, where σ is the internal stress and Y is Young's modulus of the material. Equating strains yields the thermal stress equation: σ = −Y α ΔT. For a copper bar (Y = 110 GPa, α = 16.80 × 10⁻⁶ K⁻¹) heated through ΔT = 50 °C, the resulting compressive stress is σ = 110 × 10⁹ Pa × 16.80 × 10⁻⁶ K⁻¹ × 50 K = 92.4 MPa (924 atmospheres). If the bar has a cross-sectional area of 1.0 cm² (1.0 × 10⁻⁴ m²), the compressive thrust against the end walls reaches F = σ A = 9,240 newtons (~940 kg-force). This dramatic thermal stress explains why railway tracks require expansion gaps or prestressed continuous welding, why concrete bridges feature expansion joint combs, and why dilatometers must permit one end of the specimen to expand freely against the measuring instrument.

A principal source of systematic error in jacketed steam dilatometry is longitudinal temperature non-uniformity along the specimen rod. In an ideal dilatometer, saturated dry steam at 100 °C flows rapidly through the jacket, bathing the entire surface of the rod and establishing an isothermal boundary condition. However, at the two extremities where the rod emerges through rubber stoppers or glands to contact the fixed stop and micrometer spindle, significant conductive heat loss occurs along the exposed metal ends to the ambient room air. Furthermore, if condensed water pools along the bottom of the jacket, the lower surface of the rod may experience a different heat transfer coefficient than the upper surface exposed to turbulent steam vapor. The effective thermal elongation of the rod represents the integral of the local thermal strain across the full length: ΔL = ∫_0^{L_0} α [T(x) − T_0] dx = L_0 α (T_avg − T_0), where T_avg is the true length-averaged temperature. If the thermometer bulb is located near the central steam inlet, it may record 100.0 °C while the exposed rod ends are at only 85 °C to 90 °C, causing T_avg to be lower than the recorded temperature T_therm. Neglecting this end-loss gradient causes the calculated expansion coefficient α = ΔL / [L_0(T_therm − T_0)] to be systematically underestimated by 1% to 3%. To minimize this error, well-designed dilatometers utilize thermal insulating collars around the projecting tips and maintain high steam throughput.

The precision micrometer screw gauge is the primary displacement sensor in the dilatometer apparatus. A standard metric micrometer features a precision ground lead screw with a pitch of p = 0.500 mm, meaning one complete 360° revolution of the thimble advances the spindle by exactly half a millimeter. With 50 equal graduations etched along the circumference of the thimble, the least count (smallest directly readable division) evaluates to LC = pitch / divisions = 0.500 mm / 50 = 0.010 mm (10 microns). With careful interpolation using an optical loupe or vernier sleeve graduations, readings can be estimated to ±0.001 mm (1 micron). However, screw mechanisms are vulnerable to mechanical backlash: the slight clearance between the male screw threads and female nut threads required to permit low-friction rotation. If the screw is reversed during a measurement, the spindle does not move immediately until the thread clearance is traversed. To eliminate backlash error, the micrometer must always approach the contact point from the same direction (advancing forward toward the rod). Furthermore, the zero reading R_0 must be recorded under identical directional approach at baseline temperature before heating.

The magnitude of the linear thermal expansion coefficient α varies widely across engineering materials, reflecting differences in crystal structure, bond strength, and melting points. Refractory metals with deep interatomic potential wells and high melting points generally exhibit small expansion coefficients: tungsten (T_m = 3,422 °C, α ≈ 4.5 × 10⁻⁶ K⁻¹), molybdenum (α ≈ 5.0 × 10⁻⁶ K⁻¹), and platinum (α ≈ 9.0 × 10⁻⁶ K⁻¹). Common engineering metals possess moderate expansion rates: iron and mild steel (α ≈ 11.5–12.0 × 10⁻⁶ K⁻¹), copper (α ≈ 16.80 × 10⁻⁶ K⁻¹), brass (α ≈ 19.0 × 10⁻⁶ K⁻¹), and aluminum (α ≈ 23.1 × 10⁻⁶ K⁻¹). The difference between copper and steel underlies the operation of bimetallic strips: when bonded together and heated, the copper expands more than the steel, forcing the strip to bend into an arc toward the steel side, widely exploited in mechanical thermostats, fire alarms, and thermal circuit breakers. In 1896, Charles Édouard Guillaume discovered Invar (Fe-36% Ni), an alloy possessing an extraordinarily low expansion coefficient: α_invar ≈ 1.2 × 10⁻⁶ K⁻¹ near room temperature. In Invar, an anomalous spontaneous volume magnetostriction arising from the alignment of ferromagnetic d-electron spins expands the lattice as temperature drops, almost exactly canceling the normal thermal contraction from lattice vibrations over the range −20 °C to +100 °C, earning Guillaume the 1920 Nobel Prize in Physics.