Activity and the Activity Coefficient of the Solvent

Pour ethanol into water, and neither molecule feels quite the same environment it had in the pure liquid. Hydrogen bonds rearrange, local packing tightens, and the vapor pressure that the solvent exerts no longer obeys the simple proportionality of Raoult's law. Chemists needed a way to keep the clean mathematical form of ideal-solution thermodynamics while quietly absorbing all the messy non-ideality into one correction factor. That factor is the activity coefficient of the solvent. It matters because almost every real solution used in the laboratory, from seawater to a fermentation broth, deviates from ideality. Without activities, equilibrium constants, boiling-point elevations, and phase diagrams would all carry unpredictable errors. This article builds the Raoult's-law-based (Convention I) definition of solvent activity from the chemical potential. You will see how the activity coefficient is measured from vapor pressure. You will work two full numerical examples. Finally, the article foreshadows how the same logic extends to solutes and to the Gibbs–Duhem relation between the two.

Key takeaways

Solvent activity $a_1$ is defined so that $\mu_1 = \mu_1^{*} + RT\ln a_1$ , mirroring the ideal-solution equation exactly.
Convention I (Raoult's law basis) sets $a_1 \to x_1$ as $x_1 \to 1$ , using the pure liquid as the reference state.
The activity coefficient $\gamma_1 = a_1/x_1$ measures deviation from Raoult's law; $\gamma_1 = 1$ for an ideal solution.
Experimentally, $a_1 = p_1/p_1^{*}$ , so vapor pressure measurements give activities directly.
Positive deviations ( $\gamma_1 > 1$ ) mean weaker solvent–solute interactions than solvent–solvent; negative deviations mean stronger ones.
Key idea: replacing mole fraction with activity, $x_1 \to a_1 = \gamma_1 x_1$ , lets every ideal-solution equation remain valid for real solutions, with all the non-ideal physics packed into $\gamma_1$ .

The core insight

The solvent activity coefficient is nothing more than a bookkeeping device that rescales mole fraction so the exact chemical potential equation keeps the algebraic form of the ideal solution, $\mu_1 = \mu_1^{*} + RT\ln a_1$ , while $\gamma_1 = a_1/x_1$ carries all the real physics of solvent–solute interactions and is measured directly from the ratio of actual to ideal vapor pressure.

The physical picture

In an ideal solution, the solvent's chemical potential follows directly from Raoult's law: the vapor pressure above the solution, $p_1$ , is proportional to the mole fraction, $p_1 = x_1 p_1^{*}$ , where $p_1^{*}$ is the vapor pressure of the pure solvent. Substituting into the equilibrium condition between liquid and vapor gives the familiar result $\mu_1 = \mu_1^{*} + RT\ln x_1$ . This is elegant, but it holds only when solvent molecules in solution experience essentially the same intermolecular environment as in the pure liquid — true only in the dilute limit or for chemically similar components. Real solutions break this proportionality. Ethanol–water mixtures show positive deviations (the vapor pressure is higher than Raoult's law predicts) because hydrogen bonding is disrupted upon mixing. Chloroform–acetone mixtures show negative deviations because a new, stronger hydrogen bond forms between unlike molecules. Rather than abandon the clean logarithmic form of the chemical potential, chemists define an "effective mole fraction" — the activity $a_1$ — that reproduces the true chemical potential exactly:

\mu_1 = \mu_1^{*} + RT\ln a_1

The activity coefficient $\gamma_1$ is then defined by $a_1 = \gamma_1 x_1$ , so that all the non-ideality is isolated in $\gamma_1$ , a single measurable number for each composition and temperature. This is Convention I, sometimes called the Raoult's-law-based or "solvent" convention, because the reference state is the pure liquid solvent itself: as $x_1 \to 1$ , we require $\gamma_1 \to 1$ and hence $a_1 \to x_1$ . This convention is what preserves the ideal-solution equations in form for the solvent, and it is the natural choice for the majority component in a mixture. A different convention (Henry's law basis) is used for solutes, treated in the companion article.

Figure

Vapor pressure of the solvent versus mole fraction: ideal line and two real deviations

Figure

Molecular picture of the reference state and the activity correction

The core relations

\mu_1 = \mu_1^{*} + RT\ln a_1

where $\mu_1^{*}$ is the chemical potential of the pure liquid solvent at the same $T$ and $p$ , and $a_1$ is the solvent activity.

a_1 = \gamma_1 x_1, \qquad \gamma_1 \to 1 \text{ as } x_1 \to 1

where $\gamma_1$ is the (Convention I, Raoult's-law-based) activity coefficient of the solvent.

a_1 = \frac{p_1}{p_1^{*}}

where $p_1$ is the equilibrium partial vapor pressure of solvent above the solution and $p_1^{*}$ is the vapor pressure of the pure solvent.

\gamma_1 = \frac{p_1}{x_1\, p_1^{*}}

where the right side is the experimentally accessible ratio of observed to Raoult's-law-predicted vapor pressure.

\mu_1 - \mu_1^{*} = RT\ln x_1 + RT\ln \gamma_1

where the first term is the ideal (entropic, mixing) contribution and the second is the excess (non-ideal) contribution.

The core relations, continued: excess chemical potential

\mu_1^{\text{E}} = RT\ln\gamma_1

where $\mu_1^{\text{E}}$ is the solvent's excess chemical potential, linking activity coefficients directly to excess Gibbs energy (see Real Solutions and Excess Functions).

Derivation AFrom vapor–liquid equilibrium to the solvent activity coefficient

Goal. Starting from the equality of chemical potentials at vapor–liquid equilibrium, derive $a_1 = p_1/p_1^{*}$ and show that this reduces correctly to Raoult's law in the ideal limit.

Step 1. Equilibrium condition. At equilibrium, the chemical potential of the solvent must be equal in the liquid and vapor phases:

\mu_1(\text{liquid}) = \mu_1(\text{vapor})

Treating the vapor as an ideal gas (a good approximation at modest pressures), the vapor-phase chemical potential is

\mu_1(\text{vapor}) = \mu_1^{\ominus}(\text{g}) + RT\ln\!\left(\frac{p_1}{p^{\ominus}}\right)

where $p^{\ominus}$ is the standard pressure and $\mu_1^{\ominus}(\text{g})$ is the standard chemical potential of gaseous solvent.

Step 2. Apply the same relation to the pure liquid. For the pure liquid solvent in equilibrium with its own vapor at pressure $p_1^{*}$ ,

\mu_1^{*}(\text{liquid}) = \mu_1^{\ominus}(\text{g}) + RT\ln\!\left(\frac{p_1^{*}}{p^{\ominus}}\right)

Both expressions share the same gas-phase standard state, so we can eliminate $\mu_1^{\ominus}(\text{g})$ by subtraction.

Step 3. Subtract to eliminate the gas standard state. Subtracting the pure-liquid equation from the solution equation:

\mu_1(\text{liquid}) - \mu_1^{*}(\text{liquid}) = RT\ln\!\left(\frac{p_1}{p^{\ominus}}\right) - RT\ln\!\left(\frac{p_1^{*}}{p^{\ominus}}\right)

The standard pressure cancels inside the logarithm difference:

\mu_1 - \mu_1^{*} = RT\ln\!\left(\frac{p_1}{p_1^{*}}\right)

Step 4. Match to the definition of activity. Comparing this to the defining relation $\mu_1 = \mu_1^{*} + RT\ln a_1$ term by term forces the identification

a_1 \equiv \frac{p_1}{p_1^{*}}

This is an operational definition: activity is measured, not merely postulated, because $p_1$ and $p_1^{*}$ are both directly observable vapor pressures.

Step 5. Recover Raoult's law as the ideal limit. If the solution happens to be ideal, experiment shows $p_1 = x_1 p_1^{*}$ for all $x_1$ . Then

a_1 = \frac{x_1 p_1^{*}}{p_1^{*}} = x_1

so $\gamma_1 = a_1/x_1 = 1$ identically, confirming consistency with the ideal-solution result derived in Ideal Solutions and the Chemical Potential of Liquids.

Step 6. Isolate the activity coefficient. For a real solution, dividing the operational activity by the mole fraction gives the working formula used in every vapor-pressure experiment:

\boxed{\gamma_1 = \frac{p_1}{x_1\, p_1^{*}}}

Any measured deviation of this ratio from unity is, by construction, entirely attributable to non-ideal solvent–solute interactions.

Worked examples

Example 1 — Foundational: computing $a_1$ and $\gamma_1$ from vapor pressure data. At 25 °C, pure water has vapor pressure $p_1^{*} = 3.17$ kPa. A solution of a nonvolatile solute has mole fraction of water $x_1 = 0.900$ , and the measured vapor pressure of water above this solution is $p_1 = 2.85$ kPa. Step 1: Compute the activity directly from the vapor pressure ratio.

a_1 = \frac{p_1}{p_1^{*}} = \frac{2.85\ \text{kPa}}{3.17\ \text{kPa}} = 0.899

Step 2: Compute the activity coefficient.

\gamma_1 = \frac{a_1}{x_1} = \frac{0.899}{0.900} = 0.999 \approx 1.00

Step 3: Interpret. Since $\gamma_1 \approx 1$ , this solution is very nearly ideal at this composition — a reasonable result, since many dilute aqueous solutions of small nonvolatile, non-dissociating solutes (such as sucrose) obey Raoult's law closely near $x_1 = 0.9$ . Step 4: Compute the excess chemical potential contribution.

\mu_1^{\text{E}} = RT\ln\gamma_1 = (8.314\ \text{J K}^{-1}\text{mol}^{-1})(298.15\ \text{K})\ln(0.999)

\mu_1^{\text{E}} = (2478.8\ \text{J mol}^{-1})(-0.0010) = -2.5\ \text{J mol}^{-1}

Sanity check: a near-zero excess chemical potential is consistent with $\gamma_1 \approx 1$ ; the magnitude (a few joules per mole) is tiny compared to typical $RT\ln x_1$ terms (here $RT\ln(0.900) \approx -261$ J mol⁻¹), confirming the ideal-mixing term dominates.

Example 2 — Applied: ethanol–water, a solution with genuine positive deviation. At 25 °C, pure ethanol has vapor pressure $p_1^{*} = 7.87$ kPa. In a solution with ethanol mole fraction $x_1 = 0.400$ (a water-rich mixture), the measured ethanol partial vapor pressure is $p_1 = 5.93$ kPa (values representative of real ethanol–water vapor–liquid equilibrium data at this composition). Step 1: Compute what Raoult's law alone would predict, for comparison.

p_1^{\text{ideal}} = x_1 p_1^{*} = (0.400)(7.87\ \text{kPa}) = 3.148\ \text{kPa}

Step 2: Compute the actual activity from the measured vapor pressure.

a_1 = \frac{p_1}{p_1^{*}} = \frac{5.93\ \text{kPa}}{7.87\ \text{kPa}} = 0.7535

Step 3: Compute the activity coefficient.

\gamma_1 = \frac{a_1}{x_1} = \frac{0.7535}{0.400} = 1.884

Step 4: Interpret the size of the deviation. Since $\gamma_1 = 1.88 \gg 1$ , ethanol molecules escape the solution far more readily than Raoult's law predicts. This reflects disruption of the water hydrogen-bond network by ethanol's hydrophobic ethyl group, raising ethanol's effective escaping tendency — the classic signature of positive deviation and the origin of the ethanol–water azeotrope near $x_1 \approx 0.89$ (95% ethanol by mass). Step 5: Compute the excess chemical potential.

\mu_1^{\text{E}} = RT\ln\gamma_1 = (8.314)(298.15)\ln(1.884)

\mu_1^{\text{E}} = (2478.8\ \text{J mol}^{-1})(0.6335) = 1570\ \text{J mol}^{-1} = 1.57\ \text{kJ mol}^{-1}

Sanity check: this excess chemical potential is large and positive, over 1.5 kJ mol⁻¹, three orders of magnitude bigger than Example 1's near-ideal case — consistent with the well-documented strongly non-ideal behavior of ethanol–water mixtures, and with $\gamma_1$ noticeably exceeding 1.

Limiting cases & common traps

Situation	Behavior of $\gamma_1$	Physical meaning
$x_1 \to 1$ (solvent-rich limit)	$\gamma_1 \to 1$ by definition	Convention I reference state; recovers Raoult's law
Ideal solution at all $x_1$	$\gamma_1 = 1$ everywhere	No excess Gibbs energy; solvent–solvent and solvent–solute forces identical
Positive deviation	$\gamma_1 > 1$	Solvent–solute interactions weaker than solvent–solvent; escaping tendency enhanced
Negative deviation	$\gamma_1 < 1$	Solvent–solute interactions stronger (e.g., new H-bonds); escaping tendency suppressed
Dilute solute limit for the solvent	$\gamma_1 \to 1$ faster than solute's $\gamma_2$	Solvent obeys Raoult's law even while solute obeys only Henry's law

The chemist behind it

The activity concept traces to Gilbert N. Lewis, who in the 1900s–1920s sought a way to preserve the mathematically convenient forms of ideal thermodynamics — dilute-gas-like logarithmic terms — for real, imperfect systems. Rather than rederive every equation of state from scratch for each real substance, Lewis proposed replacing concentration or pressure with a corrected "effective" quantity: the activity, and later the fugacity for gases. This turned non-ideality into a single measurable correction factor rather than an obstacle to using the elegant ideal formalism. François-Marie Raoult's earlier vapor-pressure law (1880s), which defines the reference behavior activities are measured against, is the other essential pillar; Lewis's insight was to make Raoult's proportionality the anchor rather than an approximation to abandon.

Origins · 1 minGilbert N. Lewis and how this was discovered

Gilbert N. Lewis

1875–1946 · American chemist who introduced fugacity and activity, unifying real and ideal thermodynamic descriptions

Ask the tutor: How does the Gibbs–Duhem equation force a mathematical relationship between the solvent's activity coefficient and the solute's, so that you cannot choose them independently across the whole composition range?

Why do we bother defining activity instead of just using real vapor pressure data directly?

Activity lets every equation derived for ideal solutions — chemical potential, equilibrium constants, colligative property formulas — remain valid in form for real solutions. You substitute $a_1$ for $x_1$ everywhere. This means chemists do not need a separate theoretical framework for every non-ideal system; one activity coefficient per component captures all the extra physics, and it can be tabulated, fit to models like the Margules or van Laar equations, or predicted from group-contribution methods.

Is the activity coefficient always constant for a given solution?

No. Since $\gamma_1$ depends on composition, it generally changes with $x_1$ , along with temperature and pressure. Convention I guarantees only the boundary condition $\gamma_1 \to 1$ as $x_1 \to 1$ . Away from that limit, $\gamma_1$ can be greater than, less than, or (rarely) equal to 1 across the whole composition range; it is a function, not a fixed number, and its full composition dependence is usually reported as a table or fitted curve.

How is this different from the activity coefficient used for the solute?

The solvent's Convention I anchors $\gamma_1 \to 1$ at $x_1 \to 1$ , using the pure liquid as reference. The solute typically uses Convention II (Henry's-law basis), anchoring $\gamma_2 \to 1$ as $x_2 \to 0$ , the infinite-dilution limit, because pure solute may not even share the same phase behavior as its dilute solution. Both conventions are explored together via the Gibbs–Duhem equation, which links how $\gamma_1$ and $\gamma_2$ must vary together across the full composition range.

Can gamma1 ever be measured without vapor pressure data?

Yes. Freezing-point depression, osmotic pressure, and electromotive force measurements in electrochemical cells can all yield solvent activities through their own equilibrium relations, each reducing to the same $\mu_1 = \mu_1^{*} + RT\ln a_1$ framework. Vapor pressure is simply the most direct and commonly taught method because it connects to Raoult's law transparently, but activities are a thermodynamic state function, obtainable from any consistent equilibrium measurement.

What causes gamma1 to be less than 1 rather than greater than 1?

Negative deviations occur when solvent–solute attractive interactions exceed the solvent–solvent interactions in the pure liquid, most often through new or stronger hydrogen bonding or dipole interactions upon mixing. The classic textbook example is chloroform and acetone, where a weak hydrogen bond forms between the acidic C–H of chloroform and the carbonyl oxygen of acetone, lowering the vapor pressure below the Raoult's law line and giving $\gamma_1 < 1$ .