Steam tables

Learning zone

Why one number is enough on the saturation line

Water boiling in an open pot is at 100 °C and at atmospheric pressure, and it is not free to be at anything else. That is the rule the whole steam trade is built on. Along the saturation line, where liquid and vapour sit in equilibrium, pressure and temperature are not independent variables. Fix one and the other is decided, and so is every other property of both phases.

Gibbs’ phase rule says why. For a single component in two phases, the number of independent intensive variables is F = C − P + 2 = 1 − 2 + 2 = 1. One degree of freedom. That is exactly why a steam system is controlled from a pressure gauge: set the pressure and you have set the temperature the coil delivers heat at, and the trap, the main and the heat exchanger can all be sized from that one number.

Step off the line and the rule changes. Superheated steam and compressed liquid are single-phase states, F = 1 − 1 + 2 = 2, so they need two numbers. That is not an inconvenience of the tables, it is the physics, and it is why the superheat panel on this page asks for a pressure and a temperature while the saturated one asks for either.

Why latent heat falls as pressure rises

Condensing a kilogram of steam at atmospheric pressure releases 2,256.5 kJ, about 970 BTU per pound, at constant temperature. That isothermal release is the reason to use steam at all. But the number falls the harder the system is pushed: 1,939.7 kJ/kg at 200 °C, 1,404.8 at 300 °C, and exactly zero at the critical point, 373.946 °C and 22.064 MPa.

The reason is visible in the specific volumes. Latent heat is the work of pulling molecules out of a liquid and pushing back the atmosphere to make room for the vapour, and as pressure rises the vapour is already dense. At 100 °C a kilogram of steam occupies 1.6719 m³ against the liquid’s 0.0010435, a ratio of about 1,602 to one. At 300 °C the ratio is down to about 15. At the critical point v_f = v_g exactly, there is no room to make and no work to do, so h_fg is zero and liquid and vapour have stopped being different substances.

The plant consequence is worth stating plainly. Raising boiler pressure to get a higher steam temperature buys less heat per kilogram, so mass flow has to rise faster than the temperature gain suggests. It buys something back in pipe size, as v_g collapses at the same time, but the heat per kilogram is going the wrong way and the feedwater, deaerator and condensate load all follow it.

What superheat buys, and what it costs

Superheated steam is steam heated past its boiling point at that pressure, so it is dry, single-phase and carrying sensible heat on top of the latent heat. Turbines want it. Wet steam erodes blades, and every percent of moisture at the exhaust end costs efficiency and metal, so the throttle steam is superheated far enough that the expansion finishes with acceptable dryness. Long distribution mains want it too, because a superheated main can lose heat for a while before it starts making condensate.

Heat exchangers do not want it. The superheat has to be given up before anything condenses, and dry gas is a poor heat-transfer medium: the desuperheating zone of a coil runs at a film coefficient an order of magnitude below the condensing zone, so it eats surface area and delivers little duty. That is why process steam is usually desuperheated on the way to the plant, and why a heating coil fed with superheated steam can be short of capacity while the gauge says everything is fine.

The reference state, and why h_f at 0 °C is not zero

Read the first row of the temperature table and something looks broken: the enthalpy of saturated liquid at 0 °C is -0.0416 kJ/kg, a hair below zero, and the entropy is slightly negative too. Both are correct and both are conventions.

IF97 sets internal energy and entropy to exactly zero for saturated liquid at the TRIPLE POINT, 0.01 °C and 611.657 Pa. Everything else is measured from there. Zero °C is a hundredth of a degree BELOW that reference, so its enthalpy comes out slightly negative. There is nothing wrong with a negative enthalpy: no thermodynamic table reports an absolute energy, because no experiment measures one. Only differences are physical, and every calculation an engineer performs with these tables is a difference. A flash fraction, a heat duty, a cycle efficiency, all of them subtract one enthalpy from another and the reference cancels.

The practical warning: never mix tables. A steam table on a different reference — some older ones start at 32 °F rather than the triple point — will give the same differences and different absolute numbers, and a calculation that takes h_f from one and h_g from another is quietly wrong by the offset between them. Everything on this page comes from one formulation, which is the other reason a computed table beats a transcribed one.

What IAPWS-IF97 is, and what it is not

IF97 is not an approximation of convenience. It is the INDUSTRIAL formulation published by the International Association for the Properties of Water and Steam, and it is the formulation the printed ASME steam tables and essentially every plant performance calculation since 1997 are based on. When a turbine acceptance test or a boiler efficiency guarantee is settled, it is settled with these numbers.

There is a second standard, IAPWS-95, and it is the scientific one: a single Helmholtz equation covering the whole surface, the reference every property is ultimately traceable to. It is also implicit in almost every direction a working engineer asks a question, so give it a pressure and a temperature and it has to iterate on density before it can answer. IF97 is five separate fitted equations, each explicit in the variables that region’s users actually hold, reproducing IAPWS-95 inside the tolerances a power cycle cares about and answering without iteration in four regions out of five. Speed was a stated design requirement of it, not a side effect.

The price of five equations is five seams, and the standard publishes the size of every one of them rather than hiding it. Crossing 623.15 K from the liquid region into the near-critical region steps specific volume by about two thousandths of a percent. Nothing on this page smooths that over. The engine also refuses outright outside the standard’s range instead of extrapolating: these are polynomials fitted inside a box, one of them with a term in π²⁴(τ − ½)⁵⁸, and a step outside the box does not degrade gracefully. It produces a number that looks like steam and is noise.

Absolute, gauge, and the 99.97 °C surprise

Every equation of state is written in absolute pressure. Every gauge in a mechanical room reads the difference between the system and the atmosphere around it. The tables here carry both columns for that reason, with gauge taken as absolute minus one standard atmosphere, 101.325 kPa or 14.696 psi. Real barometric pressure moves a few percent with weather and altitude, which matters most in the vacuum rows where a kilopascal is a large share of the reading.

Which brings up the number that catches people. IF97 puts the saturation pressure at 100 °C at 101.418 kPa, slightly ABOVE one standard atmosphere. Water at exactly 101.325 kPa boils at 99.974 °C, not 100. The 100 °C boiling point belongs to the old centigrade scale, which defined itself that way; the modern kelvin is defined from the triple point instead, and the boiling point became a measured quantity that landed 26 millikelvin low. The tables on this page say 99.97 °C on the atmospheric row and mean it.