Estimating Liquid Mixture Density: Checking the Additive-Volume Assumption

Mixing, blending, and balance calculations all need the density of a liquid mixture. The natural first estimate is that volumes simply add: prepare known amounts of each component and the total volume is the sum of the separate volumes. That additive-volume assumption is easy to compute and often roughly right, but it is a modeling choice, not a law of nature. Real liquids can contract or expand on mixing—water and ethanol famously shrink—so the practical questions are when the assumption is good enough and how to correct for it when it is not. This article explains what the assumption claims, how to compute its density estimate correctly, how excess volume and partial molar volumes describe the deviations, and how accurate corrected estimates can be.

What the additive-volume assumption says

The assumption treats mixing as volume-neutral: the volume of the mixture equals the sum of the volumes its pure components would occupy at the same temperature and pressure. With component masses m_i and pure densities ρ_i, the estimated density follows from the total mass divided by this ideal volume:

V_ideal = Σ m_i/ρ_i,  so  1/ρ_ideal = Σ w_i/ρ_i, where w_i is each component’s mass fraction (its mass divided by the total mass).

Equivalently, if the amounts are known as measured pure volumes, the correct ideal estimate is the volume-weighted average of the component densities, ρ_ideal = Σ φ_iρ_i. What is not correct is averaging densities linearly by mass or mole fraction: density is nonlinear in those quantities, so the arithmetic mean of two densities is generally wrong. Educational resources on liquid mixture densities, such as the DOFPro reference page on water–ethanol mixing, emphasize this distinction because density errors propagate directly into the mass-to-volume conversions used in material balances.

An illustrative calculation

As a worked example (approximate pure densities near 20 °C: water ρ ≈ 0.998 g/mL, ethanol ρ ≈ 0.789 g/mL), imagine blending 30.0 mL of water with 70.0 mL of ethanol. The masses are about 29.9 g and 55.2 g, totaling 85.2 g. The ideal volume is 100.0 mL, so the additive-volume estimate is ρ ≈ 85.2/100.0 ≈ 0.852 g/mL. Taking the plain arithmetic mean of the two densities (≈ 0.894 g/mL) would instead overestimate by roughly 5%—an error introduced before any question of non-ideality even arises. Note also that the real density of this ethanol–water blend is a little higher than 0.852 g/mL, because the true mixture volume is smaller than 100.0 mL, as discussed next.

Why volumes are not additive: excess volume

Molecules are not inert boxes. When unlike molecules mix, their sizes, shapes, and interactions change how efficiently they pack, and the mixture volume can be smaller or larger than the sum of the parts. Thermodynamics captures this with the excess volume, V^E, defined as the difference between the actual mixture volume and the ideal (additive) volume at the same temperature, pressure, and composition. A negative V^E means contraction on mixing; a positive value means expansion.

The classic demonstration is ethanol and water: as any bartender learns, mixing 50 mL of water with 50 mL of ethanol does not yield 100 mL of drink. The shortfall—typically several millilitres out of a hundred for ethanol–water near room temperature—is the excess volume. A LibreTexts data-driven exercise builds directly on such density measurements, taken for water–ethanol mixtures at 22 °C and 40 °C, to extract excess volumes and partial molar volumes. The two temperatures matter: both pure densities and V^E change with temperature, so estimates must use data matched to the conditions of interest.

Partial molar volumes: the quantities that really add

The excess volume signals that pure molar volumes are not additive. The quantities that are additive are partial molar volumes: for component i, the partial molar volume is the rate at which the total mixture volume increases as a little more of component i is added while temperature, pressure, and the amounts of the other components are held constant. Written as V = Σ n_iV̄_i, this additivity is exact for any mixture. In water–ethanol mixtures, the partial molar volume of each component can be noticeably smaller than the pure molar volume—consistent with the negative excess volume—and it varies strongly with composition. This is why a table of pure densities alone cannot fully predict mixture volumes whenever V^E is nonzero.

A practical workflow for estimating mixture density

A workable estimation procedure has five steps:

1. Compute the ideal estimate. Use 1/ρ_ideal = Σ w_i/ρ_i with pure-component densities evaluated at the mixture’s temperature and pressure.
2. Set the accuracy requirement. If the result feeds, say, a mass-to-volume conversion in a material balance where a few percent is tolerable and the components are chemically similar, the ideal estimate may suffice.
3. Correct for excess volume. Obtain V^E as a function of composition and temperature for the system, from published mixture-density data or from correlations fitted to such data, add it to the ideal volume, and recompute the density.
4. Handle multicomponent systems with binary data. Methods exist for building ternary excess volumes from binary ones; see below.
5. Validate when it matters. A single density measurement at the target conditions provides a direct check and accounts automatically for temperature.

The payoff of the correction step is well documented. In a 1982 study, Ivan Cibulka tested methods for estimating the excess volume and density of ternary liquid mixtures of non-electrolytes using only binary data, evaluated against 27 sets of experimental data from the literature (Cibulka, 1982). The mean error in the estimated excess volume was below 0.1 cm³/mol, and in the decisive majority of the data sets the maximum error in the estimated density was lower than 0.1%—a significant reduction compared with treating the ternary mixture as ideal. In other words, adding even a binary-data-based excess-volume correction can replace percent-level errors with errors at the hundredth-of-a-percent level for this class of systems.

For choosing a correlation in the first place, the estimation literature is the resource to consult: an early review by Veeranna and Rihani (Journal of Chemical & Engineering Data, 1980) surveyed methods for estimating the densities of saturated liquid mixtures, and reviews of this kind remain useful for matching a method to a family of liquids and to the input data available.

Comparing the three levels of effort

When the assumption is safe—and when it is not

Additive volumes behave best when the components are chemically similar: neighboring homologs or isomers, molecules of comparable size that interact with unlike species much as they interact with themselves. In those cases the excess volume is close to zero, and the mass-fraction calculation gives a reliable number. Dilute amounts of one component in a dominant solvent also often change the volume nearly ideally.

The assumption is least trustworthy for water mixed with hydrogen-bonding organic liquids—alcohols are the canonical case—because strong water–alcohol attractions let unlike molecules pack more closely than like ones, producing negative excess volumes that peak at intermediate compositions. Large differences in molecular size and shape also promote deviations. Whenever such chemistry is present, or whenever the downstream calculation is sensitive to volume, use an excess-volume correction or a measurement; for well-behaved, similar liquids, the ideal estimate is usually adequate.

Common mistakes to avoid

• Averaging densities linearly. The arithmetic mean of the component densities—by equal weight, by mass fraction, or by mole fraction—is not the ideal mixture density; density enters the additive-volume formula reciprocally, as Σ w_i/ρ_i.
• Using mismatched temperatures. Pure densities and V^E both vary with temperature; water–ethanol exercises at 22 °C and 40 °C exist precisely because the deviations change with temperature.
• Treating water–alcohol and other aqueous systems as ideal. These are among the most strongly non-additive common liquid pairs.
• Quoting uncorrected estimates too precisely. An additive-volume result for a non-ideal system is typically accurate only to the percent level; report it accordingly or add the correction.
• Misreading systematic deviations as bad data. A consistent, composition-dependent gap between ideal and measured volumes is the excess volume—a property of the mixture—not experimental scatter.

Practical takeaways

The additive-volume assumption is a legitimate first estimate, provided it is implemented correctly on a mass or pure-volume basis and at matching temperature. Its reliability depends on chemistry: near-ideal for similar molecules, percent-level or worse for water–alcohol-type pairs. The structured fix is to add the excess volume, using experimental or correlation data; for ternary non-electrolyte systems, binary data alone supported density estimates with maximum errors below 0.1% in most of the data sets Cibulka tested in 1982. When accuracy matters—or when the system is water plus an associating organic—close the loop with a direct density measurement at the target conditions.