One workspace for thermophysical property data: collect components and
experimental data from many sources, compare, estimate and regress, then
hand the result to CHEMCAD or Aspen Plus. Every value stays traceable
to the method or source that produced it.
See the CHEMCAD or Aspen Plus values next to measured data and independent estimates.
Fill gaps without data
Group-contribution and predictive models estimate what nobody has measured.
Fit mixture parameters in minutes
Bring in literature data and regress NRTL, UNIQUAC, Wilson or electrolyte NRTL
against it, in Rareytec Thermo or in your linked Aspen Plus.
Send it straight back
Write the result into your CHEMCAD or Aspen Plus simulation, every value traceable.
New to the abbreviations (BIP, NRTL, UNIFAC, σ-profile, γ∞)?
The glossary explains them in one line each.
119 M
PubChem structures you can search for components
53,000
COSMO σ-profiles (CHAOS)
6
predictive mixture models
6
activity models to regress
2
simulators: CHEMCAD · Aspen Plus
Projects: start, organise, carry
A project is your workspace. It holds the components,
mixtures, data and every choice you make, and you fill it in whichever
way fits your work.
Fill it from anywhere
Components
Search by name, CAS number, formula or structure,
backed by PubChem, or draw the structure; an
unknown one becomes a user component with estimated properties.
Mixtures and commercial products stay pseudo-components; you can
confirm a representative structure so estimations run.
From a simulator
CHEMCAD: import a .ccsim or neutral
file (components and parameters), or launch Rareytec Thermo from
inside a running CHEMCAD and its components come across.
Aspen Plus: import a .bkp,
.apw, .apwz or .inp, or read the
simulation open on your PC. Components, pure-component and NRTL
parameters are checked, and every record left out is listed with the reason.
From data files and documents
PPDS/IK-CAPE files, ThermoML
archives, DDB electrolyte exports and
Reaxys query files from your own Reaxys licence.
A GC-MS hit-list PDF: the peaks are read and identified.
A publication PDFexperimental:
Claude extracts the data tables and you review them before import.
From the literature banks
Import all literature data in one run: every pair,
ternary and larger subsystem of the project's components from your own
DDB (VLE, LLE, hE, γ∞, SLE, densities, mixture heat
capacities and excess cP, critical loci, …), KDB, and
pure-component data from the NIST WebBook for the components you tick.
Progress and a Stop button while it runs, then a report of what was
imported and what could not be (with the reason). A source that is
complete is greyed out with its date, and becomes active again when
you change the components.
Keep it organised
Folders and subfolders, with the created and last-modified date of every project.
Fork a project to try another option on the same
data; both projects show where they came from.
Ready-made example projects to copy and explore.
Remove any data set you imported, and re-import it cleanly.
Take it with you: .rtp project files
One file holds the complete project: components,
chosen parameters, fitted sets, the Claude chat with its figures, and
attached files.
Encrypted (AES-256-GCM) with the Rareytec Thermo key and,
optionally, your company's own key.
Download a copy, move a project to your PC (it is
removed here), open it again when needed (it exists on
Rareytec Thermo only while open), or import it as a Rareytec Thermo project.
A CHEMCAD .ccsim can carry the project file, so a project
travels inside your simulation.
Pure components
Each component page shows the full CHEMCAD parameter set,
every value with its provenance. What the page shows is exactly what every
other page, the assistant and every export use: one set of values, used everywhere.
Choose the value you trust
For every slot, choose between your imported data, the company data
bank, the shared core bank, an estimate or your own entry.
Until you choose, a rulebook default applies: the
estimate judged best for that component (IAPWS for water, GRAPPA or
Rarey-Nannoolal, depending on the molecule).
Water from the IAPWS formulations: IF97 with its
supplementary releases, the 2008 viscosity and, below 0 °C, supercooled
water (G12-15).
Parameter source per component: This project,
a library project (reuse parameters you fitted elsewhere,
linked and write-protected instead of copied), or the CHEMCAD
databank (used for calculation, never displayed or exported).
Data, correlations and comparison
Load data and parameters from KDB and the
NIST WebBook, and boiling and melting points from PubChem.
For the Dortmund Data Bank you see what data exist (a free,
IP-free availability search). DDB licence holders import from their
own DDB through the DDB bridge.
Regress temperature-dependent data with a choice of
equations. Liquid density, viscosity, thermal conductivity and heat
capacity include high-pressure data through a Tait term.
With a linked simulation, compare against CHEMCAD's own
curves for every temperature-dependent property.
Find similar compounds by structure, or by how a
compound interacts with a diverse set of solutes and solvents.
Estimation methods
When no measurement exists, Rareytec Thermo estimates from the
structure alone. Group assignments (UNIFAC, Modified UNIFAC, PSRK) are derived
automatically and shown on the component page.
Group contribution
Rarey-Nannoolal: normal boiling point, vapour
pressure, critical data (Tc, Pc, Vc),
liquid viscosity
Rarey-Govender: liquid thermal conductivity
Rarey-Olivier: surface tension
Moller: γ∞ of organic solutes in water
A detailed page per component shows the groups,
corrections, the complete calculation and the range of validity.
How the methods work
Neural networks
GRAPPA predicts the vapour-pressure curve and the
boiling point from the molecular graph. Benchmarked against
Rarey-Nannoolal on 10,026 compounds, it is the default where it is the
more accurate method; both are always shown.
HANNA predicts activity coefficients directly from
the structure (see Mixtures).
Both come from the group of Jirasek and Hasse
(RPTU Kaiserslautern) and run unchanged on our own server, so your
structures are never sent outside.
Mixtures
UNIFAC
Modified UNIFAC (Dortmund)
UNIFAC 2.0
Modified UNIFAC (NIST)
COSMO-SAC (dsp), with 53,000 σ-profiles from the open CHAOS
database (Jirasek, Hasse, et al.)
HANNA (neural network)
Switch between them on the same plot, and see
every model's γ∞ side by side in one table.
Mixture data & regression
Click any pair or ternary to see all available data (your
imports, the shared banks and the NIST/TRC ThermoML core archive), plotted
DDB-style by isotherm and isobar, with every data set shown.
Binary interaction parameters
Regress NRTL, QNRTL, FlexQNRTL, UNIQUAC, FlexQUAC or
Wilson simultaneously on VLE, γ∞, hE,
azeotropic, SLE and LLE data, or type in parameters you already have
and see how they perform.
Vapour-pressure correction factors (one overall, one
per data set) are fitted along with the parameters. They separate
pure-component errors from mixture behaviour and never leak into the
parameters.
Carboxylic acids are treated with chemical theory
for their vapour-phase association.
Exclude single points or whole sets, restrict the temperature and
pressure window; SLE includes the ΔCp term and the eutectic.
Overlay predicted curves and, with CHEMCAD or Aspen Plus,
the simulator's own results next to the experiment.
Or let your linked Aspen Plus simulation do the
calculating: the fit then runs on Aspen's own engine (see
CHEMCAD & Aspen Plus).
Ternary systems
All ternary data in the composition triangle.
Liquid-liquid diagrams with tie lines and plait
points, from your fitted parameters, a predictive model (UNIFAC family,
COSMO-SAC, HANNA) or CHEMCAD, as switchable layers.
Gas solubility (Henry)
Gas–solvent pairs are recognised automatically (you can override) and
open a dedicated page: P–x isotherms, apparent Henry constants and
ln H over 1/T.
A robust Henry model with real-gas fugacity, pressure and composition
corrections, including an NRTL composition model with four
temperature forms.
Like any parameter set, a fitted Henry model can be linked from a
library project.
Simultaneous regression: one fit, every kind of data
Binary pairs: VLE, γ∞, hE,
azeotropes, SLE and LLE in one objective, with a vapour-pressure
correction factor per data set.
Electrolyte activity: mean ionic activity
coefficients and osmotic coefficients together, including the osmotic
coefficients derived from measured water activities and from vapour
pressures over the solution.
Electrolyte properties: a salt's density, viscosity
and thermal-conductivity parameters are fitted over every salt–solvent
pair of the project that contains it, so the salt has one set of values.
Each point counts by its own uncertainty: an osmotic
coefficient derived from a vapour pressure is weighted by the pressure
error behind it, never above a measured value.
Electrolytes & reactions
Electrolyte systems
Rareytec Thermo adds dissociation reactions with equilibrium
constants from its reaction bank, weak electrolytes such as
carboxylic acids and ammonia included, and you add the solid salts that
may crystallise.
Components with ionisable groups but no stored reaction are flagged,
so no speciation is missed unnoticed.
Models: eNRTL (CHEMCAD form), ElecNRTL
(Aspen form), a thermodynamically consistent ElecNRTL
(Rareytec Thermo) and LIFAC, predictive for salts in
organic solvents.
Each salt–solvent pair opens a hub: its parameters,
model estimates, a table of every kind of data the banks hold for it, and
all data in one spreadsheet download.
Regress the activity model
Fit the eNRTL interaction parameters (with a 1/T term over a wide
temperature range) to γ± and osmotic coefficients at once.
Osmotic coefficients are also derived exactly from measured water
activities and from vapour pressures over a non-volatile salt.
Standard errors for every parameter, outlier rounds,
a temperature and pressure window, and right-click exclusion of single
points on every diagram.
Fit with Rareytec Thermo's own engine, or with Aspen Plus
ELECNRTL in your linked simulation; the parameters carry Aspen's
names (GMELCC, GMELCD, …) either way.
Reactions
Equilibrium and kinetic reactions, shown as balanced equations.
Keq(T) calculated from formation data and heat capacities.
Kinetics exported as a ready-to-use VBA unit operation for CHEMCAD.
Electrolyte properties, transport included
Density, viscosity, electric and thermal conductivity, refractive
index, surface tension, speed of sound and diffusion each get their
own page per salt–solvent pair, with every data set. As in Aspen Plus and CHEMCAD,
you choose the model for each property once for the whole project, and it runs
through the data point by point with its deviations per data set.
Fitted models
Density: Clarke (Aspen VAQCLK, CHEMCAD Clark
correction), fitted on densities and apparent molar volumes together, or
Debye-Hückel volume (Aspen VAQDH) for aqueous solutions.
Viscosity: Jones-Dole (Aspen MUL2JONS), on the
IAPWS 2008 viscosity of water.
The transport models take their concentration from the project's
density model. When the density fit changes, every fit that depends on it
is marked stale until you refit.
Fitted parameters are stored per salt for the whole project.
CHEMCAD & Aspen Plus
Each project targets one simulator. Everything you selected
(constants, correlations, binary interaction parameters, reactions) goes back
into it, and the simulator's own calculations come to Rareytec Thermo.
CHEMCAD
Calculate
Link a simulation on the server: pure-component and
mixture curves come straight from CHEMCAD's thermo engine.
Or use your own CHEMCAD: a small
bridge on your PC runs
VLE, hE and γ curves in your licensed copy. Point it at a
.ccsim on your PC, or start Rareytec Thermo from inside a
running flowsheet. The simulation never leaves your PC; only the computed
numbers travel.
Move and export
Move the whole project: Return to CHEMCAD, Save simulation
to Rareytec Thermo, Load onto this PC.
Export a CHEMCAD neutral file or write directly into a
.ccsim, kinetics included.
Aspen Plus
Link and calculate
Link an Aspen Plus simulation on the server: binary
isobars, isotherms and hE come straight from Aspen's own
property engine and appear next to the experimental data.
Fit with Aspen's engine
The linked simulation's engine calculates every data point, so the
fitted parameters are right for exactly the model Aspen uses, whatever
options the simulation has set.
Find parameters lists the parameters the simulation
uses and probes which of them your data can determine. The fit reports
standard errors, and you keep or discard the result.
Works for vapour pressures, binary pairs and electrolyte pairs
(ELECNRTL). Trial values go into the engine's memory, never into your
file, and a fit is marked stale when the linked file changes.
Import and export
Import from Aspen files, or from the simulation open on your PC through
the Aspen bridge, which
only reads: it never saves, runs or closes your simulation.
Parameters shown in Aspen's own names and units (PLXANT, CPIGDP, …).
Export an Aspen input file (.inp, PROP-DATA) with
constants, correlations and NRTL parameters, or a fresh
.bkp.
Your licence, your PC
The CHEMCAD, Aspen and DDB bridges run on your own computer,
listen only there (127.0.0.1) and answer only to this site. Rareytec Thermo
server never connects to your PC, and nothing runs without your click.
Operated by you, or by Claude
Switch to Claude mode at the top and you are talking to an AI that knows
Rareytec Thermo, operates every page for you, and writes code on the fly for
calculations and plots that are not built in. One example of a single request:
“I react hexanoic acid with propylene oxide. Add the
starting materials and the first 100 oligomers to the project, estimate their
properties and UNIFAC groups, add the reactions with one adjustable rate
constant, and export everything as a CHEMCAD simulation.”
Claude adds all 102 components, estimates the properties,
assigns the groups, sets up the reactions, and you save the result as a
.ccsim, kinetics included.
Looks up real numbers in Rareytec Thermo's own data first,
and searches the literature on request, with sources
cited. For a confidential project it asks before searching.
Attach a publication PDF in the chat and its data go
through the reviewed import.
Figures appear in the chat, ready to download for a presentation.
Takes a restore point before every change to your project.
The chat, figures included, is kept with the project and travels with
its .rtp file.
Claude is enabled per account. Ask us to switch it on for yours.
What you can rely on
Traceable
Every value carries the source or method that produced it, down to the
reference behind an imported number.
Exact or nothing
A component is identified only on an exact match. If a name is
ambiguous, Rareytec Thermo asks instead of guessing.
Your data stays yours
Projects are private; simulations stay on your PC; neural-network
models run on our own server.
Nothing silent
When a record is skipped, a value is out of range or a calculation
fails, you are told why.