Skip to content

Utilities

Chemistry Tools includes a few standalone utilities alongside the main retrosynthesis/forward-synthesis workflow — useful on their own, or as a quick side-check while working through a route in Interactive Planner or Tree Builder.

Access

Chemistry Tools requires an active subscription to any subject — see Chemistry Tools Overview for the full access model.

Solubility Prediction

A machine-learning solubility estimator (built on the model of Vermeire, Chung, and Green) that predicts solid solubility given just the SMILES of a solute and solvent, and a target temperature.

How it works

  • The model's baseline prediction is solubility in water at 298 K. From there, it adjusts to your chosen solvent using predicted solvation free-energy differences between water and that solvent.
  • If you already know the solute's solubility in some other solvent, you can supply that as a reference point — doing so generally improves the accuracy of the prediction for your target solvent.
  • For temperatures other than 298 K, the model adjusts using a thermodynamic cycle built from separate ML predictions of sublimation enthalpy, solvation enthalpy, and phase-specific heat capacities.

Using it

  1. Enter the solute and solvent as SMILES.
  2. Enter the target temperature.
  3. Optionally supply a known solubility value in a different solvent, if you have one, to anchor the prediction.
  4. Run the prediction — for screening multiple solvents at once, batch input is supported so you don't have to repeat the same solute against each solvent one at a time.

Limitations to keep in mind

  • The model is more reliable at predicting relative trends between solvents (which solvent dissolves the compound better) than at predicting precise absolute solubility values — use it to rank candidate solvents, not as a guaranteed number.
  • Coverage is limited to neutral solvents in the liquid phase and neutral solutes in the solid phase — ionic species and other phases aren't supported.
  • For target temperatures above roughly 350 K, accurate predictions are only available for the subset of solvents (roughly 100) with well-documented critical properties.

This tool is most useful for narrowing down solvent choices during route planning — for example, screening a handful of candidate recrystallization solvents before picking one to actually try.

Buyable Compounds

The "$" markers you see on chemical nodes in Interactive Planner and Tree Builder come from a buyable-compounds database — a large catalog of commercially available starting materials cross-referenced against multiple vendor sources, including eMolecules, Sigma-Aldrich, LabNetwork, Mcule, ChemBridge, and Chemspace, together covering several hundred thousand purchasable compounds.

A route whose leaf nodes are all flagged buyable is one where every starting material can, in principle, actually be purchased rather than needing to be synthesized itself — which is exactly what makes a synthetic route "complete" in Tree Builder's results.

Admin-managed, not user-managed

Buyable-compound data is uploaded and maintained by the ASKCOS deployment admin (canonical SMILES, price-per-gram, vendor source, and lead time, with optional metadata like vendor links or availability status) — this isn't something you configure per search. If you believe a compound you know is commercially available isn't showing as buyable, contact support@examcatalyst.ai.

Troubleshooting

Symptom Likely cause
Solubility prediction seems off for your solvent Check the "relative trends, not absolute values" note above — try comparing a few candidate solvents against each other rather than trusting one number in isolation.
A compound you can buy isn't flagged "$" The buyable-compounds database may not include that specific vendor/listing — this doesn't mean the compound can't be sourced, only that it isn't in this catalog.

Still stuck? See Chemistry Tools Overview for general troubleshooting, or email support@examcatalyst.ai.