Skip to main content
v2026.11,986 entries · CC-BY 4.0

CoolProp: Open-Source Fluid Property Library Explained

CoolProp is a free C++ fluid-property library. Learn how PropsSI works, which backends exist, how it differs from REFPROP and how to cite it.

Written and maintained by CASRAI Editorial Board

Published

On this page

CoolProp is a free, open-source C++ library that calculates thermodynamic and transport properties of fluids (density, enthalpy, entropy, saturation temperature, viscosity and similar quantities) from an equation of state, and it can be called from Python, MATLAB, Excel and many other environments. Give it a fluid and two known state variables, and it returns the property you ask for in SI units. Its project site states that it may be used for commercial projects and academic work free of charge.

Definition and scope

The CoolProp project home page describes the software as a C++ library of fluid properties. It provides Helmholtz-energy equations of state and transport properties for pure and pseudo-pure fluids (the home page, read on 2026-11-12, gives a count of 136), mixture calculations built on Helmholtz formulations, incompressible fluids and brines, humid-air routines, the IAPWS-IF97 formulation for water and steam, the cubic equations of state Soave-Redlich-Kwong and Peng-Robinson, and PC-SAFT. It also offers tabular interpolation methods for speed and describes a user-friendly interface to the NIST REFPROP database.

The reference paper is Bell, Wronski, Quoilin and Lemort, “Pure and Pseudo-pure Fluid Thermophysical Property Evaluation and the Open-Source Thermophysical Property Library CoolProp”, Industrial & Engineering Chemistry Research 2014, 53(6), 2498-2508, DOI 10.1021/ie4033999 (open-access copy in PubMed Central, PMC3944605). The paper states that the most accurate thermodynamic properties are obtained from multiparameter Helmholtz-energy-explicit formulations, which is the model family CoolProp is built around.

CoolProp is a property calculator, not a process simulator. It does not solve flowsheets or unit operations; it supplies the fluid property values that such calculations need. For flowsheet-level work see the guide to Aspen Plus chemical process simulation, and for the wider physical background see thermal sciences.

How it works in practice

The PropsSI call

The high-level interface is a single function, PropsSI(output, key1, value1, key2, value2, fluid). The documentation example PropsSI('T','P',101325,'Q',0,'Water') returns the saturation temperature of water at 1 atm, in kelvin. All inputs and outputs are in SI units (K, Pa, J/kg, mol/m³ for molar quantities), so a laboratory pressure in bar or a temperature in degrees Celsius must be converted before the call.

State variables and input pairs

For a pure fluid, two independent state variables fix the state. Saturation states are specified by pairing pressure or temperature with the vapor quality Q, where Q = 1 is saturated vapor and Q = 0 is saturated liquid. Constants such as the critical temperature are obtained with dummy arguments, for example PropsSI("Tcrit","",0,"",0,"Water"). The documentation notes that the choice of input pair affects calculation speed: temperature and density (T, rho) is the fastest, pressure and temperature (P, T) is roughly 3 to 10 times slower, and pressure and enthalpy (P, H) is much slower still.

Choosing a backend

The default backend is HEOS, the Helmholtz equation-of-state implementation. A prefix on the fluid name selects another backend: REFPROP:: routes the calculation to NIST REFPROP (which must be installed separately), INCOMP:: selects incompressible fluids and brines (the documentation example is INCOMP::MEG-20%), and IF97:: uses the IAPWS-IF97 formulation, which is faster than the default but valid for water only. Mixtures are written into the fluid string with mole fractions, for example REFPROP::R32[0.697615]&R125[0.302385]. The data table below compares the five backend options and the claims each source makes about them.

Wrappers and where you can call it from

The project home page lists two tiers of language support. The distinction matters because a high-level wrapper exposes the property functions but not the full object-oriented interface.

Support level Environments listed on the project home page (2026-11-12)
Full-featured wrappers Python, C++, Modelica, Octave, C#, VB.NET, MathCAD, Java, Android, MATLAB
High-level interface only LabVIEW, EES, Excel, LibreOffice, JavaScript, PHP, Fortran, Maple, Mathematica, Scilab, Delphi/Lazarus, Julia

The source code is on GitHub under the MIT license, with a “Cite this repository” entry provided by the project.

CoolProp and REFPROP side by side

REFPROP is the NIST Reference Fluid Thermodynamic and Transport Properties Database, Standard Reference Database 23. NIST describes Version 10 as covering 147 pure fluids, 5 pseudo-pure fluids (such as air) and mixtures of up to 20 components, with 121 predefined mixtures. It is authored by Lemmon, Bell, Huber and McLinden and is distributed as a Windows program with FORTRAN subroutines and data files, plus a graphical interface and an Excel sample. One of its authors, Bell, is also the lead author of the CoolProp paper, and CoolProp can call REFPROP through the REFPROP:: prefix.

Question CoolProp NIST REFPROP
What it is Open-source C++ library with many language wrappers NIST Standard Reference Database 23, Windows program plus FORTRAN subroutines
Fluids (as stated by each project) 136 pure and pseudo-pure fluids, plus incompressible fluids and humid air 147 pure fluids, 5 pseudo-pure fluids, mixtures up to 20 components
Licensing statement MIT license; free for commercial and academic use Distributed by NIST as a database product (see the NIST SRD page for current terms)
Interoperability Can call REFPROP if it is installed Called from CoolProp, from Excel sample or from FORTRAN

The two counts are not directly comparable, because each project defines and counts fluids in its own way and both change between releases.

Where it matters for labs and process work

Property libraries sit under many calculations: energy balances on heat exchangers, refrigerant cycle analysis, steam and water properties, and brine or glycol coolant properties. The documented features map onto these needs directly. The INCOMP:: backend covers coolant-type fluids such as the MEG-20% example, the humid-air routines cover psychrometric work, and IF97:: serves steam and water calculations. The 2014 paper also reports why tabular interpolation exists: for subcooled water, the TTSE and bicubic interpolation methods were more than 120 times faster than evaluating IAPWS-95 directly (over 98% less time), and bicubic interpolation was generally orders of magnitude more accurate than TTSE for R245fa. This matters when a property call sits inside a loop such as a dynamic simulation or an optimisation, where call speed can dominate run time. See also the CASRAI guide on mass transfer for a neighbouring topic where property data feed the calculation.

A checklist of pitfalls drawn from the documentation

  • Convert to SI first. Pressure in Pa, temperature in K, energy in J/kg. The documented inputs and outputs are SI, so convert bar and Celsius values before calling.
  • Use a valid pair. Two state variables for a pure fluid; pair P or T with Q for saturation states.
  • Expect trouble at saturation. Within about 1E-4 % of the saturation pressure, both liquid and vapor are defined and the call may return an error. The documentation lets you force the phase with a suffix such as T|liquid.
  • Do not impose the wrong phase. A wrongly imposed phase yields errors or wrong values, not a warning that you chose badly.
  • Derivatives are single-phase only. The exception is saturation derivatives at Q = 0 or Q = 1.
  • REFPROP must be installed. The REFPROP:: prefix does nothing useful on a machine without it.
  • Use IF97 only for water. It is faster for water but is not a general backend.

Common misunderstandings

“CoolProp cannot do mixtures”

The 2014 paper states that mixtures were not yet handled and that work was ongoing. The current project site documents mixture calculations built on Helmholtz formulations, so the older statement is out of date. Fluid and wrapper counts have also moved: the paper reported 110 pure and pseudo-pure fluids plus 40 incompressible fluids and humid air at publication, against 136 pure and pseudo-pure fluids on the home page today.

“All properties are equally accurate”

Not according to the paper. Thermodynamic properties come from high-accuracy Helmholtz formulations, but only 36 fluids had fluid-specific viscosity and thermal-conductivity correlations at publication; the remaining fluids used extended corresponding states. The transport side was less mature than the thermodynamic side. Check the documentation for the specific fluid before relying on a viscosity or conductivity value.

“Pseudo-pure means a made-up fluid”

A pseudo-pure fluid, such as air, R404A, R410A, R407C, R507A or SES36 in the 2014 paper, is a mixture treated as if it were a single fluid. For these, the paper says saturation pressures come from ancillary equations.

“Free means no citation needed”

The project asks users to cite its work in publications. The 2014 paper (DOI 10.1021/ie4033999) is the reference to cite, and the GitHub repository provides a citation entry for the software itself.

FAQ

Is CoolProp free to use?

Yes. The project home page says it can be used for commercial projects and academic work free of charge, and the GitHub repository carries an MIT license.

How do I cite CoolProp?

Cite Bell, Wronski, Quoilin and Lemort (2014), Industrial & Engineering Chemistry Research 53(6), 2498-2508, DOI 10.1021/ie4033999, as the project requests, and use the repository citation entry for the software version you ran.

What is the difference between CoolProp and REFPROP?

CoolProp is open source and has many language wrappers; REFPROP is the NIST Standard Reference Database 23, distributed as a Windows program with FORTRAN routines. CoolProp can use REFPROP as a backend if REFPROP is installed.

Which units does PropsSI use?

SI units throughout: kelvin, pascals, J/kg and mol/m³.

Why does a saturation calculation return an error?

Very close to the saturation pressure both phases are defined, and the documentation says the call may error; specify the phase explicitly or pair a state variable with Q.

Sources

Data: CoolProp property backends and what the documentation says about each

Compares the five calculation routes documented for CoolProp (HEOS, REFPROP, IF97, INCOMP and tabular interpolation) by fluid-string prefix, coverage and documented speed or installation notes, compiled from the CoolProp project documentation, the 2014 CoolProp paper and the NIST REFPROP page.

Period covered: 2014/2026.

CoolProp property backends and what the documentation says about each
BackendFluid-string prefixCoverage as documentedDocumented noteSource
HEOS (default)HEOS:: or noneHelmholtz equations of state for pure and pseudo-pure fluids, and mixtures built on Helmholtz formulationsDefault backend of PropsSICoolProp high-level API documentation; CoolProp home page
REFPROPREFPROP::Fluids and mixtures available in NIST REFPROP (SRD 23, Version 10: 147 pure fluids, 5 pseudo-pure fluids, mixtures up to 20 components)REFPROP must be installed on the machineCoolProp high-level API documentation; NIST REFPROP page
IAPWS-IF97IF97::Water and steam onlyFaster than the default backend, for water onlyCoolProp high-level API documentation
Incompressible fluids and brinesINCOMP::Incompressible fluids and brines, for example INCOMP::MEG-20%Documented in the high-level API as a separate backendCoolProp high-level API documentation; CoolProp home page
Tabular interpolation (TTSE and bicubic)Not applicable (interpolation of tabulated values)Speed-oriented interpolation of properties, tested in the 2014 paper on subcooled water and R245faTTSE and bicubic more than 120 times faster than direct IAPWS-95 for subcooled water; bicubic generally orders of magnitude more accurate than TTSE for R245faBell et al. 2014, Ind. Eng. Chem. Res. 53(6), 2498-2508

Sources

Compiled by CASRAI from the sources above. Free to reuse under CC BY 4.0 with attribution.

Follow CASRAI

Research-administration guidance, standards updates and independent tool reviews.

Ask CASRAI · free to try

Ask about CoolProp: Open-Source Fluid Property Library Explained

2 free questions a day, no account.

An AI assistant for research: funder rules, compliance, methods and publishing. It links the sources behind its answers and labels answers that come from a web search.

Answers draw on CASRAI's guides and dictionary plus the federal and funder documents we index: Federal Register, Grants.gov, Regulations.gov and UKRI.

Works on this site and inside Claude, Cursor and the AI tools you already use.

Everything CASRAI publishes — this page, the dictionary, the guides and the news — stays open to read, with no account and no card.

Referenced across the research world

University of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logoUniversity of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logo
  • University of Cambridge logo
  • Columbia University logo
  • Crossref logo
  • University of Edinburgh logo
  • Harvard University logo
  • University of Oxford logo
  • Princeton University logo
  • Stanford School of Medicine logo
  • University College London logo
  • ORCID logo

View CASRAI adoption →