Ask what temperature a solvent boils at and the honest answer is a question in return: at what pressure? A boiling point is not a property of a liquid the way its molar mass is. It is the temperature at which the liquid’s saturated vapor pressure happens to equal the pressure above it, which makes it a property of the system — and in a rotary evaporator the pressure is a dial you control. This article gives the numbers you need for the twelve solvents that dominate laboratory practice, the equation those numbers come from, an interactive calculator for any pressure in between, and the reason the Δ20 °C rule works.
In one paragraph: the boiling point at any pressure follows from the Antoine equation, log₁₀(P/mmHg) = A − B / (T/°C + C), whose constants are substance-specific and valid only within the temperature interval over which they were fitted. Because vapor pressure rises steeply and non-linearly with temperature, the first reduction in pressure buys the largest temperature drop and each further reduction buys less. Solvents respond very differently: to boil at 40 °C, dichloromethane needs essentially no vacuum, ethanol needs about 178 mbar and water about 74 mbar. The familiar Δ20 °C rule — coolant 20 °C below the vapor, bath 20 °C above it — is not folklore but a statement about three thermal resistances in series. And a chart of pure-solvent values is a starting point, not a setpoint: mixtures, dissolved solids and real process conditions all move the number. For the full technique, see the complete guide to rotary evaporation.
Why a boiling point is a system property
Boiling begins when the saturated vapor pressure of a liquid equals the pressure of its surroundings. At sea level that pressure is about 1013 mbar, and the temperature at which each solvent reaches it is the number printed on the bottle. Reduce the surrounding pressure and the same equality is satisfied at a lower temperature, so the solvent can boil before the bulk liquid reaches its normal boiling point, reducing the thermal exposure required for solvent removal.
That single mechanism is why rotary evaporation exists as a distinct technique, and it matters far beyond convenience. Thermally driven separations dominate the energy budget of chemistry — a widely cited analysis in Nature placed chemical separations at roughly 10–15% of global energy consumption1 — and at the bench the same economics appear as thermal exposure rather than kilowatt-hours. Solvents also dominate the mass balance: process mass intensity was adopted across the pharmaceutical industry because it makes that dominance visible,2 and the E-factor literature identifies solvent as the largest single contributor to waste per unit product.3 Every solvent removal is therefore both a thermal decision and a materials decision, and the pressure setting is where both are made.
The Antoine equation and its limits
Here P is the saturated vapor pressure, T the temperature, and A, B, C are substance-specific constants. The equation is empirical rather than derived: its authority rests on how well it reproduces measured data within a stated range, a point made comprehensively in the classical review of the form and its fitting.4 Two practical consequences follow, and both are routinely ignored.
Constants are unit-dependent. The convention used above expresses pressure in mmHg and temperature in degrees Celsius. Other compilations use bar, kPa or Pa with kelvin, and their constants are numerically different. Mixing constants between conventions produces confident-looking numbers that are simply wrong — one of the most common silent errors in this calculation.
Constants are valid only over the interval they were fitted to. Extrapolating below the lower bound of a coefficient set is not a small approximation; it is an unsupported claim. High-accuracy modern work extends the range with Wagner-type correlations fitted to ebulliometric measurement — the vapor pressures of acetonitrile, for example, have been determined from 4.3 kPa upward and correlated that way, with a separate Antoine form quoted for the sub-atmospheric region that laboratory evaporation actually occupies.5 For bench purposes the Antoine form is entirely adequate, provided the validity range is respected, which is why the tool below refuses to extrapolate beyond it and says so rather than guessing.
Interactive calculator
Choose a solvent and drag the pressure. The curve is the vapor-pressure line; the horizontal line is the pressure you have set; their intersection is the boiling point. The two guide lines mark the bath and coolant temperatures suggested by the Δ20 °C rule discussed below.
Model: the Antoine equation in the mmHg / °C convention stated above, converted to mbar for display, evaluated only within each solvent’s published coefficient validity range. Where the computed temperature falls outside that range the tool reports it rather than extrapolating; where the result is valid but outside the plotted axis it says so separately. Property values are approximate reference data, not manufacturer data. A schematic teaching aid, not performance data for a specific system.
Pressure charts for common solvents
Pressure required for a 40 °C boiling point
Forty degrees is a common teaching target rather than a universal compatibility limit: warm enough to evaporate briskly with a 60 °C bath, and low enough for many thermally sensitive residues. Reading the required pressure at a fixed boiling point is the quickest way to see how far apart solvents really are.
| Solvent | Pressure for a 40 °C boiling point | Normal boiling point |
|---|---|---|
| Dichloromethane | ~1006 mbar | 40 °C |
| Acetone | ~566 mbar | 56 °C |
| Chloroform | ~474 mbar | 61 °C |
| THF | ~402 mbar | 66 °C |
| n-Hexane | ~373 mbar | 69 °C |
| Methanol | ~351 mbar | 65 °C |
| Ethyl acetate | ~251 mbar | 77 °C |
| Acetonitrile | ~230 mbar | 82 °C |
| Ethanol | ~178 mbar | 78 °C |
| 2-Propanol (IPA) | ~142 mbar | 82 °C |
| Toluene | ~79 mbar | 111 °C |
| Water | ~74 mbar | 100 °C |
Three readings are worth pausing on. Dichloromethane needs essentially no vacuum to boil at 40 °C — its normal boiling point already is 40 °C, which is why DCM comes off so readily and why it is more difficult to capture than less volatile solvents under the same coolant conditions. Toluene and water land close together at 79 and 74 mbar despite normal boiling points 11 °C apart. And the spread across the cabinet is more than an order of magnitude, from about 1000 mbar down to about 74 mbar, which is precisely why a single habitual pressure setting cannot serve every solvent.
The same solvents at 30 °C and 50 °C
Forty degrees is a convention, not a requirement. Thermally fragile material often wants 30 °C; a robust residue with a high-boiling solvent may tolerate 50 °C and finish considerably faster. The steepness of the vapor-pressure curve means those ten-degree shifts move the required pressure by much more than ten percent.
| Solvent | For 30 °C | For 40 °C | For 50 °C |
|---|---|---|---|
| Acetone | ~380 mbar | ~566 mbar | ~819 mbar |
| Ethyl acetate | ~160 mbar | ~251 mbar | ~380 mbar |
| Ethanol | ~104 mbar | ~178 mbar | ~294 mbar |
| 2-Propanol (IPA) | ~81 mbar | ~142 mbar | ~239 mbar |
| Toluene | ~49 mbar | ~79 mbar | ~123 mbar |
| Water | ~42 mbar | ~74 mbar | ~123 mbar |
Two patterns are worth extracting. across this range the required pressure rises by roughly a factor of two to three for a 20 °C increase, depending on the solvent — a useful mental check when a number looks wrong. And the ordering by bottle label is not the ordering by required vacuum: at 50 °C toluene and water both sit near 123 mbar despite normal boiling points 11 °C apart. Ranking solvents by their atmospheric boiling points does not rank them by the vacuum they need.
Using the chart in practice
The practical sequence is short. Decide the highest temperature the residue tolerates and subtract about 20 °C to get a target vapor temperature. Read the required pressure for that vapor temperature. Set the coolant about 20 °C below the vapor temperature and check that the condenser can carry the resulting latent load. Then start the run and treat all three numbers as a first setting rather than a specification — the condensation front on the condenser is the real feedback, and it will tell you within minutes whether the cooling side is keeping up.
Reading the curve rather than a single point
Because vapor pressure rises steeply and non-linearly with temperature, the relationship between pressure reduction and temperature reduction is not proportional. Going from 1013 to 500 mbar buys a large drop in boiling point; going from 100 to 50 mbar buys much less. The practical implication is that the first, easy stage of pressure reduction delivers most of the thermal benefit, and that chasing the last few tens of millibar may buy only a small additional temperature reduction while increasing volumetric vapor flow through the pump and vapor path. Condenser duty rises only if the mass evaporation rate rises.
Where the Δ20 °C rule comes from
The best-known heuristic in the field — often stated as 20/40/60, coolant at 20 °C, vapor at 40 °C, bath at 60 °C — sounds arbitrary until it is read as a statement about heat flow. There are three thermal resistances in series, and each needs a temperature difference to drive it.
Heat must pass from the bath through the flask wall and the liquid film into the boiling liquid. The vapor must then travel to the condenser and give up its latent heat through the condenser wall into the coolant. Too small a bath-to-liquid difference and the evaporation rate collapses. Too small a vapor-to-coolant difference and vapor escapes uncondensed into the pump, where it is lost material, a pump problem and an exhaust problem at once. Roughly 20 °C on each side is the empirical compromise that keeps both driving forces adequate without pushing the sample hotter than necessary.
The middle term is where the film physics enters. In a thin film heat crosses the liquid mainly by conduction and the resistance scales with thickness, so a rotating flask — which spreads a thin, continuously renewed film over a large fraction of the wall — makes the bath-to-liquid resistance small. The dynamics and stability of such films constitute a substantial physics literature,6, 7 and the engineering question of how much heat a thin falling film actually transfers has been reviewed critically for industrial evaporator geometries.8, 9 Those studies provide a useful qualitative analogy, although the geometry and hydrodynamics differ from a rotating flask.
Treat Δ20 as the first setting and the observed condensation front as the real feedback. Foaming systems want a smaller bath differential; high-boiling residues and viscous concentrates may need a larger one; a condenser working near its capacity limit needs a larger coolant differential, not a smaller one. Note also that a chiller’s capacity is specified at a stated coolant temperature, and the capacity available at your intended setpoint must be verified from published model data or confirmed with the manufacturer.
Four ways the pressure setting goes wrong
Set too high. The solvent boils sluggishly or not at all, and the usual reaction is to raise the bath — which defeats the purpose of the technique and exposes the residue to exactly the temperature the vacuum was meant to avoid. If the rate is disappointing, check the pressure before touching the bath dial.
Set too low. Boiling becomes vigorous, the vapor load rises above what the condenser can absorb, and solvent passes to the pump. The symptom is liquid accumulating downstream and a receiving flask that holds less than the flask lost. Deeper is not better once the condenser is the limiting subsystem.
Reached too fast. A rapid pull-down on a full, cold flask leaves the liquid well above its new boiling point before nucleation begins; the release, when it comes, is a bumping event. Ramp the pressure rather than dropping it, and establish rotation first.
Held fixed for the whole run. As the volatile fraction is stripped away the remaining liquid boils higher, so a setting that was correct at the start becomes marginal later. Step the pressure down, or use a programmed ramp, and expect the last portion of a mixed-solvent run to behave differently from the first.
Why the chart is not a setpoint
Every number above is for a pure solvent. Real flasks rarely contain one, and three effects move the answer.
Mixtures boil between their components, and the boiling point rises as the run proceeds. In an ideal mixture each component contributes vapor pressure in proportion to its mole fraction, so the vapor is enriched in the more volatile component. As that component is stripped away, what remains is less volatile and boils higher. A run that starts comfortably at 150 mbar can stall at the same setting an hour later — not because anything failed, but because the liquid is no longer the liquid you started with. A stepped or programmed pressure reduction is the normal answer.
Azeotropes do not obey the chart at all. At an azeotropic composition the equilibrium vapor and liquid compositions are equal at the specified pressure. Pulling deeper vacuum is not a reliable general solution: some azeotropes are pressure-sensitive, while others require an entrainer, a different separation method or a change of process design. Residual water carried through from an aqueous workup is the common practical instance.
Nonvolatile dissolved solutes can raise the boiling point. A concentrated extract behaves differently from the fresh solution, and the effect grows as the run proceeds and the residue becomes more concentrated. This is one more reason the end of a run behaves differently from its beginning.
What the numbers do not tell you
They do not tell you what leaves with the solvent. Concentration is not selective. A study of metabolomic extracts concentrated on a rotary vacuum evaporator demonstrated the unavoidable loss of non-dissociating volatile metabolites such as acetone and ethanol, showed that retention of dissociating species such as organic acids depends on solution pH and buffer content, and traced how precipitating buffer salts at the end of drying trap residual analytes.10 A lower boiling point reduces thermal damage; it does not prevent volatile loss.
They do not tell you whether the solvent is a good choice. Solvent selection and handling are consistently identified as the dominant environmental levers in fine-chemical and pharmaceutical processes,11 and mature guides now rank solvents on safety, health and environmental criteria.12, 13, 14 Those guides have been surveyed and compared,15 extended to less classical and bio-derived solvents under criteria aligned with the Globally Harmonized System,16 and set against broader life-cycle frameworks;17 the field has repeatedly cautioned that a solvent scoring well on one axis may score poorly on another.18, 19 Surveys of actual usage show how slowly practice follows guidance,20 and metric work continues to argue for holistic assessment.21 Where thermal recovery is unattractive, membrane approaches such as organic solvent nanofiltration are an active alternative.22
They do not tell you whether the solvent is safe to concentrate. This is the exception that outranks every convenience. Ethers such as diethyl ether, tetrahydrofuran and 1,4-dioxane autoxidize on contact with air to form hydroperoxides and peroxides, accelerated by light and heat and beginning as soon as a container is opened.23 The laboratory-safety literature classifies such materials by whether they are hazardous on storage alone or specifically on concentration,24, 25 because peroxides are far less volatile than the parent solvent and evaporation therefore concentrates them in the residue; incident accounts repeat the same sequence of an aged ether, a routine concentration and a violent failure.26 Screen according to the current safety data sheet and your institution’s procedure, respect both receipt and opening dates, never evaporate to dryness, and do not move, open or test a container showing crystals, discoloration, stratification, unusual viscosity, pressure or a stuck cap. The hazard extends beyond ethers: spontaneous acetone peroxide formation has been characterized in consumer-aged 2-propanol,27 and for bio-derived oxygenates the peroxide number has been argued as the practical safety metric.28 Substitution can help — 2-methyltetrahydrofuran has been examined as a bio-derived THF alternative29 — but it remains a peroxide-forming ether and does not remove the requirement for screening and storage control.
- ALab™ C410 Chemistry-Design Diaphragm Vacuum Pump — oil-free vacuum source, 13 mbar ultimate vacuum and 25 L/min free air, with a published pressure/flow table. Ultimate vacuum is a capability, not the pressure a run will sit at.
- ALab™ CCP5-Mini Recirculating Chiller — −20 °C to room temperature with ±0.3 °C stability.
Frequently asked questions
What vacuum do I need for ethanol?
About 178 mbar puts ethanol’s boiling point near 40 °C, which pairs with a 60 °C bath and a 20 °C coolant under the Δ20 rule. Use the calculator above if you want a different bath temperature; the required pressure changes quickly with the target.
Why does dichloromethane need almost no vacuum?
Because its normal boiling point is already 40 °C. That convenience has a cost: DCM vapor is also more difficult to capture than less volatile solvents under the same coolant conditions. Lower coolant temperature matters more for DCM than deeper vacuum does.
Can I use the same pressure for a mixed solvent?
Only as a starting point. A mixture boils between its components and the boiling point rises as the more volatile component is stripped away, so a fixed setpoint that was comfortable at the start becomes marginal later. Step the pressure down, or use a programmed ramp.
Why does the calculator refuse to give a number sometimes?
Because the computed temperature falls outside the range over which that solvent’s Antoine coefficients were fitted. Extrapolating would return a plausible-looking number with no support behind it, so the tool reports the boundary instead.
Is a deeper pump always better?
No. Below a certain pressure the extra depth buys little additional temperature reduction. It also raises the volumetric vapor flow for the same mass flow, which loads the pump and the vapor path; the condenser’s latent duty follows the mass evaporation rate, so it rises only insofar as the deeper vacuum actually increases that rate. Deeper vacuum may be required for high-boiling solvents such as DMF or DMSO when a moderate bath temperature is the target, but for common solvents the condenser, not the pump, is usually the limiting subsystem.
How accurate are these numbers?
They are approximate pure-solvent values computed with the stated Antoine convention and rounded. Purity, dissolved solids, mixtures and real process conditions all shift them. Use them to compare solvents and to choose a starting setpoint, not as guaranteed operating values.
References
ALab™ is a trademark of ACS Material, LLC. This article is provided for general educational purposes and discusses rotary evaporation and distillation of laboratory solvents in general terms; it is not a substitute for the operating manual of any specific instrument, for the safety data sheet (SDS) of any specific solvent, or for your laboratory’s own risk assessment. Boiling points, enthalpies of vaporization and other physical properties cited here are approximate reference values that vary with purity, temperature and pressure. Equipment specifications summarize the manufacturer’s current documentation and are confirmed at quotation; always refer to the model datasheet for guaranteed ratings. Any interactive tool on this page is a schematic teaching aid based on the stated model, not predictive design or selection software. Follow your laboratory’s safety procedures for vacuum operation, heated baths, peroxide-forming solvents and solvent handling. ACS Material makes no representations or warranties, express or implied, regarding suitability for any purpose, and will not be responsible for damages resulting from use of or reliance upon this information.