Joule-Thomson Effect Simulator — Throttle Expansion and Temperature Change
From isenthalpic throttle expansion of a van der Waals gas this tool computes the Joule-Thomson coefficient mu_JT = (2a/(RT) − b)/c_p, the temperature change DT ≈ mu_JT·DP and the inversion temperature T_inv = 2a/(Rb) in real time. Adjusting T, the pressure drop DP and the vdW constants a, b updates the throttle schematic (chambers coloured by temperature) and the mu_JT(T) curve, making the boundary between the T < T_inv cooling regime and the T > T_inv heating regime immediately visible.
Parameters
Temperature T
K
Pressure drop DP
atm
vdW constant a
J·m³/mol²
vdW constant b ×10⁻⁵
m³/mol
Defaults use N₂-like van der Waals constants (a=0.138 J·m³/mol², b=3.87×10⁻⁵ m³/mol), c_p=29.1 J/(mol·K) fixed. R=8.314 J/(mol·K). 1 atm = 101 325 Pa.
Results
—
Joule-Thomson coefficient mu_JT
—
Temperature change DT
—
Inversion temperature T_inv
—
Regime (T vs T_inv)
Throttle Schematic and Temperature Contrast
Left = high-pressure chamber (P_1, T_1) / centre = throttle (porous plug, capillary, expansion valve) / right = low-pressure chamber (P_2, T_2). Background colour follows temperature: blue on the right for cooling (mu_JT > 0), red for heating (mu_JT < 0).
mu_JT(T) Curve and Inversion Temperature
Horizontal axis = temperature T (K, 50–1200) / vertical axis = mu_JT (K/atm). Blue curve = (2a/(RT) − b)/c_p, dashed red horizontal = mu_JT = 0 (inversion temperature), yellow dot = current operating point. mu_JT > 0 cools, mu_JT < 0 heats.
Theory & Key Formulas
For a real gas the Joule-Thomson coefficient in an isenthalpic throttle expansion is defined as:
For a small pressure change the temperature change and inversion temperature are:
$$\Delta T \approx \mu_{JT}\cdot \Delta P, \quad T_{\mathrm{inv}} = \frac{2a}{R\,b}$$
$a$ is the strength of intermolecular attraction [J·m³/mol²], $b$ is the excluded molar volume [m³/mol] and $R=8.314$ J/(mol·K). For $T0$ (expansion cools), and for $T>T_{\mathrm{inv}}$ we have $\mu_{JT}<0$ (expansion heats). The defaults reproduce N₂-like $T_{\mathrm{inv}}\approx 858$ K and explain why pre-cooling is essential in the Linde liquefaction cycle.
What is the Joule-Thomson Effect Simulator?
🙋
An air-conditioner's expansion valve just lets refrigerant hiss through, no piston, no turbine. How does the temperature drop if no work is being done?
🎓
That is exactly the Joule-Thomson effect. When a real gas is throttled through a porous plug, capillary or expansion valve, the temperature changes even though no external work is performed; the process is isenthalpic. The defining coefficient is mu_JT = (∂T/∂P)_H, and the van der Waals approximation gives (2a/(RT) − b)/c_p. With this tool's defaults (N₂-like: T=300 K, a=0.138, b=3.87e-5, c_p=29.1) we get mu_JT ≈ 0.251 K/atm, so a 100-atm pressure drop yields DT ≈ −25.1 K of cooling. An ideal gas has H = c_p·T, which is independent of pressure, so DT = 0 — the cooling is the signature of real-gas intermolecular attraction (the a term) doing work against itself.
🙋
So if I push T to, say, 900 K, the cooling disappears? Wait, the simulator shows positive DT — the gas actually warms?
🎓
Yes — you have crossed the inversion temperature T_inv = 2a/(Rb). With the defaults T_inv ≈ 858 K. Below T_inv attraction (the a term) dominates and mu_JT > 0 (expansion cools); above T_inv excluded volume (the b term) wins and mu_JT < 0 (expansion heats). That is why hydrogen, with T_inv ≈ 202 K, actually warms when throttled at room temperature; to liquefy hydrogen you must first pre-cool below T_inv with liquid nitrogen. Dewar accomplished this in 1898 by combining the JT effect with cascade pre-cooling.
🙋
Remind me how this differs from an ideal-gas adiabatic expansion. I remember being told an ideal gas does not change temperature when throttled.
🎓
Good memory. For an ideal gas H = c_p·T depends only on temperature, so throttling at constant H leaves T unchanged: DT = 0. A non-zero JT coefficient is therefore a direct measurement of non-ideality. Light molecules such as He and H₂ have small a and b, so T_inv is very low (45 K for He, 202 K for H₂); CO₂ and steam have large a and b and T_inv is high (~1500 K for CO₂). The intuition "throttle ⇒ cool" only works at room temperature for a limited set of gases: N₂, O₂, CO₂, R32 and so on.
🙋
Why is the DP slider negative, all the way to −500 atm? Is that just notation for a pressure drop?
🎓
Right. Defining DP = P_2 − P_1, an expansion (high → low) has P_2 < P_1, i.e. DP < 0. Plugging directly into DT ≈ mu_JT·DP gives the correct sign of the temperature change, so the slider is bounded between −500 and 0 atm. A practical compressor drop of 100 atm → 1 atm corresponds to DP ≈ −99 atm. Sweep the DP slider from 0 to −100 atm and watch DT scale linearly. A Linde first-stage throttle in air liquefaction routinely operates around 200 atm → 1 atm and provides 40–50 K of DT, enough to reach the two-phase region of N₂.
🙋
The mu_JT(T) curve makes the inversion temperature really obvious — and the a, b sliders change where T_inv falls. Could you give some concrete species?
🎓
Sure. Try a = 0.366 J·m³/mol², b = 4.29e-5 m³/mol for CO₂: T_inv climbs above 2000 K. Switch to He-like a = 0.0035, b = 2.4e-5 and T_inv collapses to about 35 K — which is precisely why helium liquefaction (Kamerlingh Onnes, 1908) lagged the rest of cryogenics by decades. Once you internalise the inversion-temperature concept, cryogenics, household refrigeration, LNG plants and hydrogen-station fuelling all fall out of the same framework. That generality is what makes the JT effect such a beautiful piece of classical thermodynamics.
Frequently Asked Questions
The Joule-Thomson effect is the change in temperature when a real gas is expanded adiabatically through a throttle (porous plug, capillary or expansion valve). For an ideal gas the enthalpy depends only on temperature, so throttling at constant H gives no temperature change; for a real gas with attractive forces (vdW constant a) and excluded volume (b) the coefficient mu_JT = (dT/dP)_H is non-zero. The van der Waals approximation gives mu_JT ≈ (2a/(RT) − b)/c_p. With this tool's defaults (T=300 K, a=0.138 J·m^3/mol^2, b=3.87×10^-5 m^3/mol, c_p=29.1 J/(mol·K)) we obtain mu_JT ≈ 0.251 K/atm; a 100-atm pressure drop yields DT ≈ -25.1 K. This is the basic principle of the Linde liquefaction cycle and of the expansion valve in a household air-conditioner.
The inversion temperature T_inv = 2a/(Rb) is the boundary at which mu_JT changes sign. For T < T_inv the attractive term 2a/(RT) dominates, mu_JT > 0 and expansion cools the gas. For T > T_inv the excluded-volume term b dominates, mu_JT < 0 and expansion heats the gas. With this tool's defaults T_inv ≈ 858 K. Measured values are about 621 K for N2, 764 K for O2, only 202 K for H2 and just 45 K for He. Hydrogen at room temperature actually warms on throttling, which is why pre-cooling is required for hydrogen liquefaction.
In a vapour-compression cycle the high-pressure liquid refrigerant (R32, R410A, etc.) drops sharply in temperature the moment it passes through the expansion valve or capillary tube into the evaporator — this is JT throttle expansion. The two-phase mixture in the evaporator absorbs heat from indoor air and is then re-compressed. Modern refrigerants are designed so that operation stays in the T < T_inv regime; a typical 10 → 3 atm pressure drop produces 15–30 K of cooling. Move the DP slider from 0 to -100 atm in this tool and you see the linear DT ∝ mu_JT·DP relation that engineers exploit when sizing an expansion valve.
An adiabatic expansion that performs external work (turbine or piston) follows Pv^gamma = const and cools even an ideal gas because the gas loses internal energy to the surroundings. JT throttle expansion is instead isenthalpic (H = const) and produces no external work; the cooling comes entirely from internal energy being spent against intermolecular attractions. For an ideal gas H = c_p·T is pressure-independent, so throttling gives DT = 0; non-zero mu_JT is the signature of real-gas behaviour. Linde's 1895 liquefaction cycle exploited the JT effect, while the 1902 Claude cycle added a work-extracting expander to improve efficiency.
Real-World Applications
Domestic air-conditioners and refrigerators: in a vapour-compression cycle the high-pressure liquid refrigerant (R32 at 30–40 atm) passes through an expansion valve or capillary tube into the evaporator (3–10 atm). The throttling produces a two-phase mixture and lowers the temperature to roughly −5 to −25 °C, which absorbs heat from the indoor air before the vapour is re-compressed. To approximate R32 in this tool you can push a up and b up modestly; you will see a larger mu_JT than for N₂, so a smaller pressure drop produces a larger cooling.
Natural-gas liquefaction (LNG) and air liquefaction: LNG plants compress methane (CH₄, a ≈ 0.228 J·m³/mol², b ≈ 4.28e-5 m³/mol) to 70–80 atm and throttle it down to near atmospheric pressure to reach the liquefaction temperature of −162 °C. The original Linde-Hampson cycle (1895) and modern megaplants such as RasGas in Qatar and Sabine Pass in the U.S. share the same backbone. For air liquefaction the single-stage Linde cycle is inefficient, so the 1902 Claude cycle and the Heylandt cycle add a work-extracting expander.
Cryostats and quantum-computing dilution refrigerators: superconducting qubits operate at millikelvin temperatures and rely on staged cooling — liquid nitrogen (77 K), liquid helium (4.2 K) and finally a ³He/⁴He dilution stage. Because the inversion temperature of helium is about 45 K, helium cannot be liquefied by a room-temperature JT throttle alone; it must be pre-cooled by liquid nitrogen first. Setting a ≈ 0.0035 and b ≈ 2.4e-5 in this tool reproduces a He-like T_inv ≈ 35 K and explains why Kamerlingh Onnes succeeded in liquefying helium only in 1908.
Hydrogen fuel-cell vehicles and refuelling stations: hydrogen has T_inv ≈ 202 K, so rapidly filling a 70-MPa hydrogen tank at room temperature actually warms the gas — the so-called "filling heat-up" — and can reach 80–90 °C, threatening tank strength. Standard SAE J2601 therefore mandates pre-cooling to about −40 °C. In this tool, setting T = 300 K, a = 0.0247 and b = 2.66e-5 produces mu_JT < 0 and reproduces the counter-intuitive heating of hydrogen on throttling.
Common Misconceptions and Caveats
The most common misconception is that a gas always cools when throttled. In reality the sign depends on the temperature and the species. For T > T_inv the gas heats up. Hydrogen, helium and neon are the textbook examples of room-temperature "warming" gases; liquefying them requires pre-cooling. Linde succeeded in liquefying hydrogen only because he pre-cooled it to 77 K with liquid nitrogen first. Setting T = 300 K and dropping a to about 0.025 in this tool collapses T_inv and reveals the warming branch.
A second pitfall is believing the Joule-Thomson effect performs work. In fact the process is isenthalpic and produces zero external work; the cooling comes from internal energy paid against intermolecular attraction. Do not confuse JT throttling with adiabatic expansion in a turbine. The latter extracts shaft work and cools even an ideal gas (Claude cycle, higher efficiency); the former is simpler, has no moving parts and is preferred in household air-conditioners despite its lower exergetic efficiency.
Finally, do not over-trust the van der Waals approximation. Near the saturation dome and at high pressure it loses accuracy, and equations of state such as Redlich-Kwong, Peng-Robinson or Lee-Kesler are required. The full JT formula is mu_JT = (T(∂V/∂T)_P − V)/c_p; the (2a/(RT) − b)/c_p shown here is its leading low-density term. Industrial process simulators (Aspen HYSYS, Aspen Plus) and reference databases (NIST REFPROP, GERG-2008) all use higher-order EOS for design accuracy. This tool is a teaching aid for the existence of an inversion temperature and the sign change of mu_JT, not a replacement for process-design software.