Partitioning, retention and separation efficiency

A phase preference becomes an amount only after volumes are known

A separation exploits differences in how constituents interact with phases or move through a medium. One form may favor a stationary phase, another a mobile phase; charged forms may have different velocities in a field. A useful separation requires these differences to persist under the actual composition, temperature and transport conditions.

For one specified chemical form at equilibrium, define a partition ratio K=c_s/c_m, with s denoting the chosen stationary phase and m the mobile phase. The order matters: reversing numerator and denominator replaces K by its reciprocal. This concentration ratio is not automatically a thermodynamic standard partition constant, nor the fraction of the total material in either phase.

If volumes V_s and V_m are accessible and concentrations are uniform within the model, phase amounts are n_s=c_sV_s and n_m=c_mV_m. Their ratio is k=n_s/n_m=KV_s/V_m. The mobile fraction of the total inventory is f_m=1/(1+k). This inventory relation assumes one form, linear equilibrium, unchanged phase volumes and no separate losses or reactions.

A fourfold concentration preference can give equal amounts

An original model assigns K=4.00, V_s=10.0 mL and V_m=40.0 mL, with total amount 1.00 mmol. Then k=1.00, so each phase contains 0.500 mmol. Mobile concentration is 0.0125 mol L⁻¹ and stationary concentration is 0.0500 mol L⁻¹. Their ratio is four while their amounts are equal.

The smaller favored phase offsets its greater concentration. A statement such as “four times as much is in the stationary phase” would confuse concentrations with amounts. These are theoretical equilibrium inventories, rather than a claim that every molecule follows an identical residence-time schedule.

In an ideal chromatographic residence model, the retention factor k also relates the stationary delay to the mobile passage time. If hold-up time t_M is the passage time of a suitable unretained reference, total retention time is t_R=t_M(1+k). With assigned t_M=90.0 s and k=1.00, t_R=180 s. Constant mobile velocity and negligible or corrected extra-column delay are part of that model.

Subtract common passage time before comparing retention

For two retained constituents A and B under the same fixed conditions, define adjusted retention t_R′=t_R−t_M. Order the pair so B is later, and define retention selectivity α=k_B/k_A=t_R,B′/t_R,A′. It compares their phase-dependent delays after removing the common passage contribution.

Given t_M=1.00 min, k_A=2.00 and k_B=3.00, the retention times are 3.00 and 4.00 min. Adjusted times are 2.00 and 3.00 min, giving α=1.50. The raw-time ratio 4/3≈1.333 answers a different question.

Selectivity describes a retention difference; it does not include peak width. Two peaks with α=1.50 can still overlap if transport spreads them broadly. Retention order can also change with solvent, phase or chemical form, so α belongs to the stated pair and conditions.

Efficiency concerns broadening within a constituent

A theoretical plate number N summarizes column efficiency in an appropriate peak model. For an exact Gaussian with standard deviation σ_t in time, N=(t_R/σ_t)². Plate height H=L/N, where L is column length. Smaller H at otherwise comparable conditions describes less broadening per length; these “plates” are conceptual and need not be physical compartments.

A Gaussian’s tangent-intercept baseline width w_b is 4σ_t: tangents at its two inflection points meet the baseline at t_R±2σ_t. The curve itself extends beyond those intercepts. Full width at half height w_h is 2√(2 ln 2)σ_t. An observed tailing or asymmetric peak needs an appropriate definition and model rather than treating these Gaussian equalities as universal.

For invented t_R=6.00 min, w_b=0.400 min and L=150 mm, σ_t=0.100 min, N=3600 and H=0.04167 mm. The corresponding Gaussian full width at half height is 0.23548 min. Using the half-height width as if it were the tangent baseline width would change the inferred N incorrectly. Extra-column broadening can also affect a measured width.

Preferential recovery enriches without proving purity

For extraction, a distribution ratio D can include all forms of the constituent in each specified phase. In a fixed equilibrium model, the extracted fraction is r=D V_e/(V_a+D V_e), where a and e label aqueous and extracting phases. Its value depends on phase volumes as well as D.

Suppose independent noninteracting symbolic solutes A and B each begin at 1.00 mmol in 30.0 mL aqueous phase. Give the 10.0 mL extracting phase D_A=8.00 and D_B=2.00. The extracted amounts are 0.72727 mmol A and 0.40000 mmol B. The A:B amount ratio rises from 1.00 to 1.81818, but B remains present.

This model gives enrichment and constituent-specific recoveries, not proof of a pure A preparation. Unmodelled constituents, changing speciation, adsorption or incomplete equilibration can alter both amounts. Purity also needs a specified basis: an A:B amount ratio alone is not the mass fraction of A in all recovered material.

Choose the performance quantity that answers the question

Partition preference, retention selectivity, efficiency, resolution and recovery describe different features. A favorable partition ratio does not guarantee high recovery with a tiny phase volume; narrow peaks do not guarantee different retention; an enriched fraction need not be pure.

Before comparing separations, identify the forms, phases, phase order, temperature, inventory basis and observation. Then state which quantity matters for the intended analytical or preparative task. The next lessons add finite contact balances, peak interpretation and transport limits to this framework.

Worked examples

Try it yourself

1. Keep K=4.00 and total amount 1.00 mmol, but use V_s=20.0 mL and V_m=40.0 mL. Find k and the stationary amount.

   Show the answer and reasoning

   k=2.00; stationary amount 0.66667 mmol.

   The stationary fraction is k/(1+k)=2/3. The mobile amount is 1/3 mmol.

2. For an assigned Gaussian at the same t_R=6.00 min and L=150 mm, double w_b from 0.400 to 0.800 min. Find N and H.

   Show the answer and reasoning

   N=900; H=0.16667 mm.

   σ_t doubles to 0.200 min; N falls by a factor of four and H rises by a factor of four.

3. Two constituents have different adjusted retention times. Which additional peak feature is needed to assess whether their signals overlap?

   Show the answer and reasoning

   Their widths and relevant peak shapes.

   Retention selectivity contains no width. Resolution and actual overlap depend on both separation in time and broadening.

4. The extract contains 0.72727 mmol A and 0.40000 mmol B. What claim can be made from this inventory, and what remains unsupported?

   Show the answer and reasoning

   The A:B amount ratio is about 1.81818 and B is present; complete purity and a mass fraction in all recovered material are unsupported.

   The model specifies only two amount inventories, not all contaminants, molar masses or a validated purity measurement.

What to carry forward

• Specify phase order and chemical form in a partition ratio.
• Phase amounts depend on accessible volumes as well as concentrations.
• Subtract hold-up time to compare retention delays.
• State the peak model and width convention in an efficiency calculation.
• Enrichment, recovery and purity answer different questions.