Resonance of Air Columns theory

Sound propagation in a gaseous fluid represents the transmission of longitudinal acoustic pressure disturbances governed by Euler's fluid momentum equation and the thermodynamic equation of state. Isaac Newton originally attempted to derive the speed of sound by assuming that acoustic compressions and rarefactions occur isothermally (at constant temperature), yielding the classic Newtonian velocity v_Newton = √(P / ρ). However, experimental measurements consistently exceeded Newton's prediction by approximately 15% to 20%. In 1816, Pierre-Simon Laplace resolved this celebrated discrepancy by recognizing that acoustic pressure fluctuations occur at acoustic frequencies (hundreds to thousands of cycles per second) far too rapid for thermal energy to diffuse between adjacent microscopic compression bands (which heat adiabatically) and rarefaction bands (which cool adiabatically). Because the thermal diffusion length in air at audible frequencies is orders of magnitude smaller than the acoustic wavelength, sound waves propagate under strictly adiabatic, isentropic conditions (entropy S = constant). Applying the adiabatic gas equation P V^γ = constant, where γ = C_p / C_v is the heat capacity ratio (adiabatic index), Laplace derived the true acoustic velocity formula: v = √(γ P / ρ) = √(γ R T / M). For dry air, which is predominantly composed of diatomic nitrogen (N₂) and oxygen (O₂) with 5 degrees of classical freedom, γ = 7/5 = 1.400, while the effective molar mass is M = 0.028964 kg/mol. At standard temperature and pressure (T₀ = 273.15 K, P = 101,325 Pa), Laplace's equation yields v₀ = 331.30 m/s, matching precision laboratory measurements with remarkable fidelity.

When an acoustic wave traveling inside a bounded duct strikes a boundary, the nature of the reflected wave depends entirely on the acoustic impedance mismatch between the duct and the terminating medium. Specific acoustic impedance is defined as the ratio of acoustic pressure to particle velocity: Z = p / u = ρ v. At the closed bottom of the air column, the dense water surface presents an acoustic impedance Z_water ≈ 1.5 × 10⁶ Pa s/m, which is more than 3,600 times greater than the characteristic acoustic impedance of air Z_air ≈ 415 Pa s/m. Because the water meniscus is virtually rigid compared to air, the longitudinal air molecules cannot undergo physical displacement at the water interface; thus, the water surface enforces a strict acoustic displacement node: u(0, t) = 0. By Euler's momentum equation ∂p/∂x = −ρ (∂u/∂t), a displacement node corresponds mathematically to an acoustic pressure antinode where pressure fluctuations are maximized: |p(0, t)| = 2 p_incident. Conversely, at the upper open aperture of the tube, the air column abruptly terminates into the unconfined ambient room, where the acoustic pressure must equalize with the constant ambient atmospheric reservoir (p ≈ 0); this forms an acoustic pressure node (displacement antinode) where longitudinal air molecules vibrate with maximum spatial amplitude.

In elementary physics treatments, the open termination of an acoustic pipe is often idealized as an absolute pressure node located exactly at the physical lip of the tube (L = λ/4). However, in realistic fluid mechanics, the vibrating air mass immediately inside the tube bore does not abruptly decouple from the exterior fluid at the geometric rim. Instead, the vibrating air column must impart momentum to an external 'plug' or hemispherical envelope of ambient air sitting immediately outside the opening. This moving external mass behaves as an acoustic radiation mass (inertance), loading the aperture. As a consequence, the acoustic pressure does not drop to atmospheric ambient pressure right at the rim; the zero-crossing of acoustic pressure occurs at an effective distance c outside the physical pipe lip, known as the end correction. In 1877, Lord Rayleigh derived the exact analytical solution for acoustic radiation from a circular cylindrical pipe of radius r. For a pipe terminating in an infinite rigid baffle (a flanged pipe), the radiation impedance yields an end correction of c_flanged = (8 / 3π) r ≈ 0.8488 r (refined by modern variational methods to 0.8216 r). For an unflanged, thin-walled cylindrical pipe suspended in open air (as used in standard university resonance apparatus), acoustic diffraction around the tube lip reduces the external inertial load, yielding the classical Rayleigh-Levine-Schwinger value: c ≈ 0.6133 r. For a standard 32 mm bore tube (r = 1.60 cm), the theoretical unflanged end correction is c ≈ 0.6133 × 1.60 cm = 0.981 cm.

The spatial boundary conditions of a closed-end tube—a displacement node at x = 0 (water level) and a displacement antinode at x = L + c (effective open aperture)—restrict the allowable standing wave patterns to discrete normal modes. Expressing the spatial standing wave displacement as u(x) = A sin(k x), where k = 2π / λ is the acoustic wave number, the boundary condition at the effective open end requires sin[k(L + c)] = ±1, which dictates that k(L + c) must be an odd multiple of π / 2: k(L + c) = (2n − 1) π / 2 for integer n = 1, 2, 3, ... Substituting k = 2π / λ = 2π f / v, we obtain the fundamental normal mode eigenvalue equation: L_n + c = (2n − 1) λ / 4 = (2n − 1) v / (4f). For the fundamental mode (n = 1), the air column accommodates exactly one-quarter of a sound wavelength: L₁ + c = λ / 4. For the second resonance mode (n = 2, representing the third harmonic or first overtone), L₂ + c = 3λ / 4. For the third mode (n = 3, fifth harmonic), L₃ + c = 5λ / 4. Crucially, closed pipes support only odd harmonics (1f, 3f, 5f, ...); even harmonics (2f, 4f, 6f, ...) possess a displacement node at both ends or an antinode at both ends, which violates the asymmetric boundary conditions of a closed-open pipe.

A frequent pedagogical error in introductory physics is to calculate the speed of sound from a single tuning fork measurement by assuming an idealized tube without end correction (v = 4 f L₁). In a 32 mm bore tube at 512 Hz, L₁ is approximately 15.8 cm; neglecting the 0.98 cm end correction introduces a systematic error of 0.98 / 15.8 ≈ 6.2%, which is unacceptably large for university metrology. To eliminate this systematic bias completely, experimentalists employ linear least-squares regression across variable frequencies. Rearranging the fundamental resonance condition isolates the physical tube length L₁ as a linear function of inverse frequency 1/f: L₁ = (v / 4) (1 / f) − c. By designating x = 1 / f (with units of seconds) as the independent variable and y = L₁ (in metres) as the dependent variable, the data fit the canonical linear form y = m x + b. Here, the slope m represents one-quarter of the acoustic propagation speed: m = v / 4, so that v = 4m. Crucially, the regression slope m is completely independent of the end correction c. The vertical intercept b directly yields the physical end correction: b = −c, so that c = −b. Ordinary least-squares regression thus accomplishes a powerful metrological separation: it evaluates the true propagation velocity of sound from the gradient while simultaneously measuring the boundary aperture end correction from the intercept.

Because the speed of sound in an ideal gas depends on the thermodynamic temperature T through v = √(γ R T / M), the experimental velocity v_t measured at room temperature t (in °C) must be referenced to standard temperature T₀ = 273.15 K (0 °C) for rigorous scientific comparison. Expressing absolute temperature as T = T₀ + t = 273.15(1 + t / 273.15), we can expand the square-root function via the binomial theorem: v(t) = v₀ √(1 + t / 273.15) ≈ v₀ [1 + t / (2 × 273.15)] = v₀ + [v₀ / 546.3] t. Substituting the standard IUPAC velocity of sound in dry air at 0 °C, v₀ = 331.30 m/s, yields the widely used engineering linear approximation: v(t) ≈ 331.30 + 0.606 t m/s. For example, at an ambient laboratory temperature of t = 20.0 °C, the exact thermodynamic formula yields v(20 °C) = 331.30 √(293.15 / 273.15) = 343.20 m/s, while the linear approximation yields 331.30 + 0.606(20.0) = 343.42 m/s. Once students obtain their experimental sound speed v_exp from regression, they extrapolate it back to 0 °C using v₀,exp = v_exp / √(1 + t / 273.15). Comparing v₀,exp with the accepted standard 331.30 m/s provides a rigorous test of whether deviations stem from ambient thermal gradients or genuine measurement scatter.

In a resonance tube experiment conducted over a water reservoir, the air column inside the tube is in continuous contact with a free liquid water surface, resulting in significant water vapor saturation (relative humidity approaching 100% near the meniscus). Water vapor (H₂O) has a molecular weight of 18.015 g/mol, which is substantially lighter than dry air (mean molecular weight 28.964 g/mol). When water evaporates into the column, it displaces heavier diatomic nitrogen (N₂, 28.01 g/mol) and oxygen (O₂, 32.00 g/mol) molecules under constant atmospheric pressure, reducing the effective bulk density ρ_mix of the moist gas mixture. Although water vapor is a triatomic molecule with a slightly lower adiabatic index (γ_vapor ≈ 1.33) than dry air (γ_air ≈ 1.40), the density reduction dominates, causing sound to travel slightly faster in humid air than in bone-dry air: v_humid = √(γ_mix P / ρ_mix) > v_dry. At 20 °C and 100% relative humidity, the presence of saturated water vapor increases the speed of sound by approximately 0.35% (about +1.2 m/s). Furthermore, viscous shear at the inner glass tube wall and thermal boundary-layer dissipation slightly attenuate the wave amplitude and lower the phase velocity according to Kirchhoff-Helmholtz tube absorption theory: v_tube ≈ v_free [1 − (1 / 2r) √(2η / ρ ω)], where η is dynamic viscosity and ω = 2π f. In university tubes (radius r > 15 mm), this wall-drag velocity reduction is less than 0.15%, but represents an essential consideration in precision acoustic metrology.