Engineering Article

Mastering Solvent Selection and Property Calculations

A comprehensive engineering guide to evaluating thermodynamic properties, utilizing Hansen Solubility Parameters (HSP), and determining safety criteria for process design.

Last Updated: May 21, 2026
14 min read
Reading Preferences
Table of Contents

Whether you are designing a liquid-liquid extraction unit, formulating a new coating, or optimizing an active pharmaceutical ingredient (API) crystallization process, solvent selection is a critical decision. A seemingly minor change in solvent can drastically alter reaction kinetics, downstream separation costs, and the overall environmental footprint of a plant.

Density

Needed to size pumps, vessels, and pipelines. Decreases with rising temperature due to thermal expansion.

Vapor Pressure

Governs volatility, flashing risk, and NPSH calculations. Modelled by the Antoine equation.

Viscosity

Controls pressure drop and heat transfer. Falls exponentially with temperature per the Andrade equation.

Solvent Selection Diagram showing overlap of Thermodynamics, Safety, and Solubility

Figure 1: The optimal solvent is found at the intersection of high solubility, favorable thermophysical properties, and safe environmental profiles.

Advertisement

1. The Role of Solvents in Chemical Processing

In engineering, a solvent acts as the medium for chemical reactions, separations, and heat transfer. The fundamental rule of thumb, "like dissolves like," refers to the intermolecular forces (polar, non-polar, hydrogen bonding) between the solute and solvent.

However, modern chemical engineering demands more precision than a rule of thumb. Engineers must rigorously calculate temperature-dependent properties such as vapor pressure, density, and viscosity to size pumps, heat exchangers, and distillation columns accurately.

Solvent Classes at a Glance

Before diving into the mathematics of property estimation, it helps to understand how solvents are broadly categorised. Each class has distinct polarity and hydrogen-bonding characteristics that largely determine where a solvent sits in Hansen space.

Class Representative Solvent Key Characteristic Typical Engineering Use
Polar Protic Ethanol, Water, Methanol H-bond donor & acceptor; high δH Crystallisation, fermentation, API extraction
Polar Aprotic Acetone, DMSO, DMF, NMP High δP; H-bond acceptor only Polymer dissolution, lithography, reactions
Non-Polar Hexane, Heptane, Toluene Low δP & δH; high δD Oil extraction, degreasing, rubber processing
Chlorinated DCM, Chloroform, TCE Dense; moderate polarity; poor H-bonder Extractions, metal degreasing (being phased out)
Ether / Cyclic THF, CPME, 2-MeTHF Moderate polarity; good H-bond acceptor Grignard reactions, polymer synthesis, APIs
Ester Ethyl Acetate, Butyl Acetate Moderate δP; low toxicity profile Coatings, food flavour, green extractions

Solvent Selection Flowchart

Not sure where to start? This decision flowchart maps the key engineering questions — from the nature of the solute through to safety constraints — to a recommended solvent class and, ultimately, to quantitative property verification.

START: Choose a Solvent Is the solute polar / ionic? No Non-Polar Hexane / Heptane Yes Does it donate H-bonds? No Polar Aprotic Acetone / DMSO Yes Temperature constraint below 50 °C? No Polar Protic Ethanol / Water Yes Regulated / EHS critical environment? No Low-Bp Protic Methanol / IPA Yes Green Solvent Candidate Ethyl Acetate · 2-MeTHF · CPME · Cyrene™ Verify: Calculate HSP (Ra), Density, Viscosity & Vapor Pressure at T

Figure 2: A structured decision flowchart for narrowing down solvent class before quantitative evaluation.

2. Thermodynamic Property Estimation

Pure component properties are rarely static; they fluctuate significantly with operating temperature. Using empirical correlations allows engineers to determine these properties across a wide temperature range.

Density and Viscosity Dependencies

Liquid density generally decreases with temperature due to thermal expansion, often modeled using a standard polynomial equation:

$$\rho(T) = A + B \cdot T + C \cdot T^2$$ (Eq. 1)

Conversely, dynamic viscosity ($\mu$) drops exponentially as a liquid is heated. The most common empirical model for liquid viscosity is the Andrade Equation:

$$\ln(\mu) = A + \frac{B}{T} + C \cdot \ln(T)$$ (Eq. 2)

Vapor Pressure: The Antoine Equation

Vapor pressure indicates a solvent's volatility and is essential for designing flash drums and distillation equipment. The most widely used correlation is the Antoine Equation:

$$\log_{10}(P) = A - \frac{B}{T + C}$$ (Eq. 3)

Where $P$ is the vapor pressure, $T$ is the temperature, and $A$, $B$, and $C$ are component-specific regression constants.

Watch Your Units: Antoine constants are highly unit-dependent! The most common literature sets require $T$ in Celsius (°C) and output $P$ in mmHg. Always verify the expected units before plugging constants into your simulation or spreadsheet.

Example 1: Evaluating Toluene Properties

Scenario: Toluene is being evaluated as a wash solvent at an operating temperature of 77 °F. Determine its density and viscosity for preliminary pump sizing.

  1. Density: Using standard polynomial liquid density correlations for Toluene at this temperature yields a density of 54.1 lb/ft³.
  2. Viscosity: Based on the Andrade equation for liquid viscosity ($\ln(\mu) = A + B/T$), the dynamic viscosity of Toluene is approximately 0.56 cP.
  3. Vapor Pressure: Applying the NIST Antoine constants ($A=6.95464$, $B=1344.800$, $C=219.482$ for T/°C and P/mmHg), the vapor pressure is calculated as 0.55 psia, confirming it is a relatively low-volatility solvent at ambient conditions.

Example 2: Acetone in a Chiller Loop

Scenario: Acetone is utilized as a low-temperature heat transfer fluid operating at 14 °F. Engineers need to verify fluid properties to size the circulation pump and estimate cavitation risk.

  1. Density: As the liquid cools, it becomes denser. At this temperature, the density of Acetone rises to 51.1 lb/ft³.
  2. Viscosity: Cold fluids are generally thicker. The viscosity increases slightly to 0.45 cP, impacting the frictional pressure drop in the pipes.
  3. Vapor Pressure: To ensure the pump's Net Positive Suction Head Available (NPSHa) is sufficient, the vapor pressure is checked. Using the Antoine equation, it is exceptionally low at 0.78 psia, minimizing the risk of cavitation at the pump suction.

Example 3: Ethanol Distillation Feed

Scenario: A feed stream of pure Ethanol enters a distillation column pre-heated to 176 °F. What are its thermophysical characteristics right before entering the column?

  1. Density: Due to thermal expansion at near-boiling temperatures, the density drops to roughly 45.9 lb/ft³.
  2. Viscosity: The elevated heat significantly thins the fluid, dropping dynamic viscosity down to roughly 0.43 cP.
  3. Vapor Pressure: Checking the NIST Antoine constants ($A=8.11220$, $B=1592.864$, $C=226.184$ for T/°C and P/mmHg), the vapor pressure leaps to 15.7 psia. Since this slightly exceeds standard atmospheric pressure (14.7 psia), the Ethanol will be actively flashing into vapor depending on the column's actual operating pressure.
Advertisement

3. Hansen Solubility Parameters (HSP)

To predict if a solvent will dissolve a specific solute (like a polymer or an API), engineers use Hansen Solubility Parameters. HSP divides total cohesive energy into three individual forces, creating a 3D coordinate system:

The HSP Distance ($R_a$)

The "distance" between the solvent (1) and the solute (2) in this 3D space is calculated using the following equation:

$$(R_a)^2 = 4(\delta_{D1} - \delta_{D2})^2 + (\delta_{P1} - \delta_{P2})^2 + (\delta_{H1} - \delta_{H2})^2$$ (Eq. 4)

Note the factor of $4$ in front of the dispersion term; this is empirically required to make the solubility "sphere" actually spherical in the plotted coordinate space.

4. Case Study: Selecting an Extraction Solvent

Once $R_a$ is calculated, it is compared to the solute's radius of interaction ($R_0$). The ratio of these two values is called the Relative Energy Difference (RED).

$$RED = \frac{R_a}{R_0}$$ (Eq. 5)

If $RED < 1$, the solvent will dissolve the solute. If $RED \approx 1$, partial dissolution occurs. If $RED > 1$, the solvent will not dissolve the solute.

HSP Compatibility Does Not Guarantee Easy Separation. A low RED value confirms a solvent will dissolve the solute, but it says nothing about how easily the solvent can later be separated from it. Two key pitfalls are:
  • Azeotrope formation: Ethanol and water are mutually soluble (compatible by polarity) yet form a minimum-boiling azeotrope at 95.6 wt% ethanol (bp 78.15 °C, below both pure components) that cannot be broken by simple distillation alone. Pressure-swing or molecular-sieve drying is required.
  • Partial miscibility: Some solvent pairs separate into two liquid phases at certain compositions or temperatures. Always check liquid–liquid equilibrium (LLE) data alongside HSP when designing extraction circuits.

Example 4: Dissolving Polystyrene

Scenario: An engineer needs to select a solvent to dissolve a waste stream of Polystyrene (PS) for recycling. The known HSP parameters for PS are: $\delta_D = 21.3$, $\delta_P = 5.8$, $\delta_H = 4.3$, with an interaction radius $R_0 = 12.7$ (units in MPa$^{1/2}$). Let's evaluate Toluene and Ethanol.

  1. Toluene Evaluation: Toluene's parameters are $\delta_D = 18.0$, $\delta_P = 1.4$, $\delta_H = 2.0$.
    Calculating the squared distance: $(R_a)^2 = 4(18.0 - 21.3)^2 + (1.4 - 5.8)^2 + (2.0 - 4.3)^2 = 68.2$.
    $R_a = \sqrt{68.2} = 8.26$.
    The RED is $8.26 \div 12.7 = $ 0.65. Since RED < 1, Toluene is an excellent solvent for Polystyrene.
  2. Ethanol Evaluation: Ethanol's parameters are $\delta_D = 15.8$, $\delta_P = 8.8$, $\delta_H = 19.4$.
    Calculating the squared distance: $(R_a)^2 = 4(15.8 - 21.3)^2 + (8.8 - 5.8)^2 + (19.4 - 4.3)^2 = 358.0$.
    $R_a = \sqrt{358.0} = 18.9$.
    The RED is $18.9 \div 12.7 = $ 1.49. Since RED > 1, Ethanol is a poor solvent and will not dissolve Polystyrene.

5. HSP Solvent Blending: Creating a Good Solvent from Two Poor Ones

One of the most powerful practical applications of Hansen Solubility Parameters is solvent blending. Because the HSP of a mixture are simply the volume-fraction-weighted averages of the pure component parameters (Eq. 6), two solvents that individually sit outside a solute's Hansen sphere can be combined to land right inside it.

$$\delta_{mix} = \phi_1 \cdot \delta_1 + \phi_2 \cdot \delta_2$$ (Eq. 6)

Where $\phi_1$ and $\phi_2$ are the volume fractions of solvents 1 and 2 respectively, and the equation applies independently to each of the three HSP components ($\delta_D$, $\delta_P$, $\delta_H$).

Example 5: Blending Solvents for a Target Polymer

Scenario: An engineer must dissolve a polyurethane coating resin. The resin's HSP are $\delta_D = 18.6$, $\delta_P = 10.2$, $\delta_H = 7.0$, with $R_0 = 7.0$ MPa$^{1/2}$. Two available solvents are evaluated individually — one is a clear non-solvent and the other only borderline. Can blending them optimise the result?

Individual solvent data:

  • Solvent A — Cyclohexane: $\delta_D = 16.8$, $\delta_P = 0.0$, $\delta_H = 0.2$ → $(R_a)^2 = 4(16.8-18.6)^2 + (0.0-10.2)^2 + (0.2-7.0)^2 = 4(3.24) + 104.04 + 46.24 = 163.24$ → $R_a = 12.8$, $RED = 1.82$ — Non-solvent
  • Solvent B — Acetone: $\delta_D = 15.5$, $\delta_P = 10.4$, $\delta_H = 7.0$ → $(R_a)^2 = 4(15.5-18.6)^2 + (10.4-10.2)^2 + (7.0-7.0)^2 = 4(9.61) + 0.04 + 0.0 = 38.48$ → $R_a = 6.2$, $RED = 0.89$ — Borderline solvent (RED just under 1)

Acetone is a marginal solvent on its own, but adding Cyclohexane improves the $\delta_D$ match. Evaluating a 50 / 50 vol% blend:

  1. $\delta_{D,mix} = 0.5 imes 16.8 + 0.5 imes 15.5 = 16.15$
  2. $\delta_{P,mix} = 0.5 imes 0.0 + 0.5 imes 10.4 = 5.2$
  3. $\delta_{H,mix} = 0.5 imes 0.2 + 0.5 imes 7.0 = 3.6$
  4. $(R_a)^2 = 4(16.15-18.6)^2 + (5.2-10.2)^2 + (3.6-7.0)^2 = 4(6.0) + 25.0 + 11.56 = 60.56$
  5. $R_a = 7.78$, $RED = 7.78 / 7.0 = $ 1.11 — still marginally outside. Try a 30 / 70 vol% (Cyclohexane / Acetone) blend instead:
  1. $\delta_{D,mix} = 0.3 imes 16.8 + 0.7 imes 15.5 = 15.89$
  2. $\delta_{P,mix} = 0.3 imes 0.0 + 0.7 imes 10.4 = 7.28$
  3. $\delta_{H,mix} = 0.3 imes 0.2 + 0.7 imes 7.0 = 4.96$
  4. $(R_a)^2 = 4(15.89-18.6)^2 + (7.28-10.2)^2 + (4.96-7.0)^2 = 4(7.34) + 8.53 + 4.16 = 42.06$
  5. $R_a = \sqrt{42.06} = 6.49$, $RED = 6.49 / 7.0 = $ 0.93 — Good solvent blend! The 30/70 mixture lands inside the Hansen sphere.

This iterative volume-fraction optimisation is exactly the kind of calculation where a digital tool saves significant time.

6. Safety and Environmental Considerations

Thermodynamic suitability is only half the battle. Process engineers must balance performance against environmental, health, and safety (EHS) constraints.

Green Solvent Substitution Cheat Sheet
Traditional Solvent Hazard Profile Greener Alternative
Hexane Neurotoxic, High VOC Heptane or Cyclopentyl methyl ether (CPME)
NMP / DMF Reprotoxic, high concern Cyrene™, DMSO, or Rhodiasolv® PolarClean
Dichloromethane (DCM) Carcinogenic, highly volatile Ethyl Acetate or 2-Methyltetrahydrofuran (2-MeTHF)
Benzene Known human carcinogen Toluene or Xylene
Advertisement

7. Common Mistakes to Avoid

Common Engineering Pitfalls in Solvent Work
  • Using Antoine constants in wrong units. Antoine constants are tabulated for specific T and P units that vary by source. Using constants from NIST (T/°C, P/mmHg) with T in Kelvin will produce an answer that is off by orders of magnitude. Always confirm the unit basis before applying any constant set.
  • Confusing Ra with R0. Ra is the calculated Hansen distance between a specific solvent and a solute; R0 is the empirically determined interaction radius of the solute itself. Comparing Ra to Ra from another solvent (instead of to R0) is a common source of errors in RED calculation.
  • Applying density correlations outside their valid temperature range. Polynomial density fits are regressed over a specific temperature interval. Extrapolating well beyond this range (especially near the boiling point or freezing point) can give physically unrealistic values. Always check the stated validity range of the correlation.
  • Assuming HSP guarantees easy separation. A low RED confirms dissolution but says nothing about the ease of solvent recovery. Check for azeotrope formation and liquid–liquid equilibrium (LLE) data independently.
  • Neglecting temperature dependence of viscosity for pump sizing. Viscosity can change by a factor of 5–10 across a typical operating temperature range. Sizing a pump at the cold-start viscosity without checking the hot running viscosity (or vice versa) leads to either an oversized or cavitating pump.
  • Ignoring HSP uncertainty in R0. The interaction radius R0 for polymers and APIs is often determined experimentally with limited data points. A RED of 0.95 and a RED of 1.05 are analytically different but practically within the uncertainty band — always run bench-scale dissolution tests to confirm predictions.
Key Takeaways
  • Solvent class (protic, aprotic, non-polar) is a fast first filter; the flowchart above maps solute characteristics to candidate classes before any calculation begins.
  • Density (Eq. 1), viscosity (Eq. 2), and vapor pressure (Eq. 3) are all strongly temperature-dependent and must be evaluated at actual operating conditions, not reference conditions from a datasheet.
  • Hansen Solubility Parameters quantify solubility using three independent forces; a RED number below 1.0 (Eq. 5) predicts dissolution, while RED above 1.0 predicts incompatibility.
  • Two non-solvents can be blended into an effective solvent by volume-fraction averaging their HSP to land inside the solute's interaction sphere (Eq. 6) — an iterative process best done computationally.
  • Thermodynamic compatibility (low RED) does not guarantee easy downstream separation; always verify azeotrope and LLE behaviour independently.
  • Antoine constants are unit-specific. Always confirm the T and P basis of any constant set before use (NIST values used in this article: T/°C, P/mmHg).

Conclusion

Selecting the ideal solvent requires a rigorous approach that integrates molecular theory, bulk thermodynamic calculations, and practical safety constraints. By understanding solvent class from the outset, quantifying temperature-dependent properties such as density, viscosity, and vapor pressure using empirical correlations, and applying Hansen Solubility Parameters to screen and blend candidates, engineers can make defensible, efficient, and environmentally responsible solvent choices. The common-mistakes checklist and blending methodology in this article are designed to prevent the most frequent sources of error in real plant and laboratory settings.

Instant Solvent Property Data

Skip the handbook searches. Use our Solvent Calculator to instantly calculate temperature-dependent thermodynamic properties—such as density, vapor pressure, and viscosity—for over 30 essential industrial solvents.

Open Solvent Property and Sizing Calculator
Further Reading
  1. Hansen, C.M. Hansen Solubility Parameters: A User's Handbook, 2nd ed. CRC Press, 2007. — The definitive reference for HSP theory, tabulated data, and experimental methodology.
  2. Green, D.W.; Southard, M.Z. (eds.) Perry's Chemical Engineers' Handbook, 9th ed. McGraw-Hill, 2018. — Comprehensive source of Antoine constants, density and viscosity correlations for industrial solvents.
  3. NIST WebBook — webbook.nist.gov — Free online database of thermophysical properties, Antoine constants, and phase-equilibrium data for pure components.
  4. Smallwood, I.M. Handbook of Organic Solvent Properties. Butterworth-Heinemann, 1996. — Concise physical and safety property tables for 72 common organic solvents.
  5. Sheldon, R.A. “Metrics of Green Chemistry and Sustainability.” ACS Sustainable Chemistry & Engineering, 2018. — Provides context for green solvent selection and environmental impact assessment frameworks.