Modes and Classification of Chromatography

Gas chromatography built on the partition mechanism dominates any analysis of volatile organics: environmental labs use GC (often coupled to mass spectrometry, GC-MS) to quantify pesticide residues and volatile organic pollutants at parts-per-billion levels, forensic labs use it for accelerant identification in arson cases and for blood-alcohol testing, and the petroleum industry uses simulated-distillation GC to characterize crude oil fractions by boiling-point distribution. Reversed-phase partition HPLC is the workhorse of pharmaceutical quality control, used to assay drug potency and detect degradation products in essentially every tablet, capsule, and injectable formulation released to market, typically monitored by UV-Vis absorbance detection. Ion-exchange chromatography, often run as fully automated ion chromatography with conductivity detection, is the standard method for quantifying anions (chloride, sulfate, nitrate) and cations in drinking water, industrial process streams, and boiler feedwater, and in biochemistry it separates proteins by surface charge as an early purification step, or resolves amino acids and nucleotides in metabolomics workflows. Size-exclusion chromatography underlies polymer characterization (reported as gel permeation chromatography, GPC, giving molecular-weight distributions and dispersity indices for plastics and coatings), and in the biopharmaceutical industry it is the required regulatory method for confirming that a therapeutic protein (a monoclonal antibody, for instance) has not aggregated into higher-molecular-weight species during manufacturing or storage. Planar chromatography persists wherever speed, low cost, and simultaneous visualization of many samples side-by-side matter more than instrumental precision: TLC remains a standard reaction-monitoring tool in every synthetic organic chemistry laboratory, and it retains a role in pharmacopeial identity testing and in field-deployable drug-screening kits. Each of these applications is, at bottom, a choice made along exactly the three axes developed here, and each feeds into the instrument-level detail covered separately for gas systems and for liquid systems.

Ask the tutor: How do I predict which chromatographic mode will resolve two specific real molecules I'm trying to separate?

Chromatography separates a mixture because different solutes distribute differently between two phases in relative motion: a mobile phase that moves and a stationary phase that stays put. The distribution is governed by an equilibrium distribution constant,

where $c_S$ and $c_M$ are the solute's molar concentration in the stationary and mobile phase, and larger $K$ means more time spent stationary, hence longer retention. Everything below is a different physical reason $K$ can differ from solute to solute, and a different physical form the two phases can take.

Axis 1 — mobile phase state. In gas chromatography (GC), the mobile phase is an inert gas (helium, nitrogen, hydrogen) that does not interact with the solute chemically; it merely transports vaporized analyte molecules past a stationary phase held in a column. Retention in GC is controlled almost entirely by the solute's vapor pressure and its affinity for the stationary liquid film — so GC is restricted to analytes that can be vaporized without decomposing, roughly compounds with boiling points below about 300 °C and molar masses below a few hundred grams per mole. In liquid chromatography (LC), the mobile phase is a liquid (water, methanol, acetonitrile, hexane, or mixtures, called the eluent) that actively participates in the equilibrium — competing with the stationary phase for the solute. Because the mobile phase itself carries chemical information (polarity, ionic strength, pH), LC can separate species GC cannot touch: proteins, ionic salts, polymers, and thermally fragile drugs.

Axis 2 — retention mechanism. Four mechanisms recur across both gas and liquid systems:

• Adsorption: the solute physically or chemically adsorbs onto discrete active sites on a solid surface (silica, alumina, activated carbon). Retention tracks polarity matching between solute and surface, and the equilibrium is fundamentally a surface saturation process, not a bulk dissolution.
• Partition: the solute dissolves into a thin liquid (or liquid-like, e.g., bonded alkyl chain) film coating an inert support. Retention tracks the classic "like dissolves like" solvation equilibrium, described by the same thermodynamics as liquid-liquid extraction.
• Ion-exchange: the stationary phase carries fixed ionic sites (sulfonate, quaternary ammonium) that hold a mobile counter-ion; analyte ions compete for those sites, displacing the counter-ion. Retention tracks charge, charge density, and the concentration of competing ions in the mobile phase.
• Size-exclusion (gel filtration/permeation): the stationary phase is a porous, chemically inert gel; there is no chemical affinity at all. Retention arises purely from whether a molecule is small enough to diffuse into the pore volume — small molecules explore more volume and elute later, large molecules are excluded and elute first.

Axis 3 — geometry. In column chromatography, the stationary phase is packed into (or coated onto the wall of) a tube, and the mobile phase is pumped or pressure-driven through it; detection happens at the outlet as a function of time, generating a chromatogram. In planar chromatography — paper chromatography and thin-layer chromatography (TLC) — the stationary phase is a flat sheet (cellulose fiber, or a silica/alumina layer on glass or plastic), the mobile phase climbs by capillary action, and all separated components remain visible simultaneously on the plate at the moment migration is stopped, quantified by the retardation factor $R_f$.

The mechanisms differ most sharply in how retention depends on solute size. In adsorption and partition chromatography, larger, more polarizable, or more polar molecules generally retain more strongly (more surface contact, more favorable solvation). In size-exclusion chromatography, the dependence inverts: retention decreases monotonically with molecular size, because retention there is purely entropic — governed by whether the molecule fits into the stationary phase's pores, not by any enthalpic affinity.

Partition chromatography obeys the same equilibrium that governs liquid-liquid extraction, expressed through the partition (distribution) coefficient:

where:

• $K_D$ = partition coefficient between stationary liquid and mobile phase,
• $c_S, c_M$ = solute molar concentration in stationary and mobile phase at equilibrium,
• $k$ = retention factor (capacity factor), the dimensionless quantity measured directly from a chromatogram,
• $V_S, V_M$ = volumes of stationary and mobile phase in the column.

Reading: $k$ is what you actually measure ($k = (t_R - t_M)/t_M$ from retention times), while $K_D$ is the underlying thermodynamic constant; the phase ratio $V_S/V_M$ is an instrumental property of the column, so the same stationary-phase chemistry gives different $k$ on columns with different phase ratios.

Ion-exchange chromatography is governed by a mass-action equilibrium between the analyte ion and the mobile-phase counter-ion competing for fixed resin sites. For a monovalent analyte ion $A^{+}$ exchanging with a monovalent eluent ion $E^{+}$ on a cation-exchange resin (resin sites denoted R⁻):

where:

• $K_{ex}$ = selectivity (exchange equilibrium) coefficient, intrinsic to the ion pair and resin,
• $[A^+]_S, [A^+]_M$ = analyte ion concentration bound to resin and free in mobile phase,
• $[E^+]_S, [E^+]_M$ = eluent (competing) ion concentration bound and free.

Reading: because $[E^+]_M$ appears in the denominator, raising the eluent's ionic strength (more competing counter-ion) decreases $k$ for the analyte — this is why ion-exchange separations are controlled experimentally by ramping salt concentration (a "salt gradient") rather than by changing solvent polarity as in reversed-phase LC.

Size-exclusion chromatography has no equilibrium constant at all in the adsorption/partition sense; retention is described purely by an accessible-pore-volume fraction $K_{SEC}$:

where:

• $V_R$ = retention volume (elution volume) of the solute,
• $V_0$ = interstitial (void) volume between packing particles — the elution volume of a molecule fully excluded from all pores,
• $V_i$ = total internal pore volume of the packing,
• $K_{SEC}$ = fraction of the pore volume accessible to that solute (1 for a molecule small enough to fully permeate every pore, 0 for one too large to enter any pore).

Reading: $K_{SEC}$ is a steric partition coefficient, not a chemical one — it depends only on the ratio of solute hydrodynamic radius to pore radius, which is why SEC calibration curves plot $\log(\text{molar mass})$ linearly against $V_R$ over the useful pore-size range of a given gel.

Example 1 (foundational) — partition GC: finding $K_D$ from a measured retention factor. A packed GC column contains a stationary liquid phase (polydimethylsiloxane-type) with $V_S = 0.15$ mL total film volume and a mobile-phase (carrier gas) volume inside the column of $V_M = 3.20$ mL (the column's void volume at operating temperature). A hydrocarbon solute is injected; the dead-time marker (unretained air peak) elutes at $t_M = 0.85$ min, and the solute peak elutes at $t_R = 4.62$ min. Step 1 — retention factor from the chromatogram.

Step 2 — phase ratio.

Step 3 — partition coefficient.

So the solute favors the stationary liquid phase by a factor of roughly 95 to 1 in concentration terms — a physically reasonable value for a moderately volatile hydrocarbon on a nonpolar liquid phase, consistent with $K_D$ values (10–500) typically reported for packed-column GC of C6–C10 alkanes. As a limiting-case check: if the injected compound were the carrier gas itself (or any truly unretained marker), $t_R \to t_M$, so $k \to 0$ and $K_D \to 0$ — correctly predicting no partitioning for a species that does not dissolve in the stationary liquid at all.

Example 2 (applied) — ion-exchange HPLC: predicting how a buffer change shifts retention. A cation-exchange column separates a mixture of protonated amines. At a mobile-phase (eluent) ammonium ion concentration of $[E^+]_M = 20.0$ mM, a particular analyte cation elutes with $k_1 = 8.40$. The column operator wants to know the expected retention factor if the buffer is changed to $[E^+]_M = 55.0$ mM, all else (temperature, resin, flow rate) held constant. Step 1 — apply the inverse proportionality derived above, $k \propto 1/[E^+]_M$, for the case of monovalent analyte and monovalent eluent ion:

Step 2 — solve for $k_2$.

Step 3 — convert to a predicted retention time, using a measured column dead time of $t_M = 1.35$ min:

Raising the buffer concentration by a factor of 2.75 cut the retention factor by essentially the same factor (2.75), exactly as the derived proportionality predicts, and moved the analyte off the column more than 4 minutes earlier (from an original $t_{R,1} = 1.35(1+8.40) = 12.7$ min down to 5.47 min). As a limiting-case check: if $[E^+]_M \to \infty$ (a very concentrated eluent flush), the model predicts $k \to 0$, meaning the analyte co-elutes with the void volume — physically correct, since an overwhelming excess of competing ion strips essentially all bound analyte off the resin instantly, which is exactly the mechanism used to regenerate or strip ion-exchange columns.

Example 3 (exam / competition level) — size-exclusion chromatography: resolving two proteins and predicting a third's elution volume. An SEC column is calibrated with two protein standards. Cytochrome c (12.4 kDa) elutes at $V_{R,1} = 14.8$ mL. Bovine serum albumin (66.5 kDa) elutes at $V_{R,2} = 9.6$ mL. The column's void volume (measured with blue dextran, fully excluded, $M \gg$ exclusion limit) is $V_0 = 7.20$ mL, and the total permeation volume (measured with a small molecule like sodium azide) is $V_0 + V_i = 18.4$ mL. Step 1 — find $K_{SEC}$ for each standard using $V_R = V_0 + K_{SEC}V_i$, first computing $V_i = 18.4 - 7.20 = 11.2$ mL. For cytochrome c:

For BSA:

Step 2 — build the calibration line. SEC theory (and near-universal empirical practice) uses a linear relationship between $K_{SEC}$ and $\log_{10}(M)$ over the working range of the gel:

Using the two calibration points, $\log_{10}(12{,}400) = 4.093$ and $\log_{10}(66{,}500) = 4.823$:

Step 3 — predict the elution volume of an unknown protein of $M = 35{,}000$ g/mol.

The predicted elution volume (11.6 mL) falls correctly between BSA's 9.6 mL (larger protein, elutes earlier) and cytochrome c's 14.8 mL (smaller protein, elutes later), consistent with the monotonic inverse-size ordering that defines SEC. As a limiting-case check: setting $M \to \infty$ in the calibration equation would force $K_{SEC} \to -\infty$, which is unphysical — correctly signaling that the linear log-M calibration is only valid within the gel's working pore-size range; above the exclusion limit every large molecule genuinely converges on the same value, $K_{SEC}=0$ and $V_R = V_0$, regardless of how much bigger it is.

The classification scheme described here rests on foundational work by Archer John Porter Martin and Richard Laurence Millington Synge, who in the 1940s developed partition chromatography and, crucially, the theoretical plate model that lets chemists reason quantitatively about any of these mechanisms using the same retention-factor and efficiency language. Martin, a British biochemist working initially on problems in wool and amino-acid analysis, recognized that a separation could be engineered by controlling a solute's partition between two immiscible liquid phases held stationary and mobile, rather than by the traditional single-stage extraction. Synge, his collaborator, contributed the rigorous physical-chemical treatment that connected column behavior to distillation theory, borrowing the "theoretical plate" concept directly from industrial fractionation columns. Their 1941 paper on partition chromatography of amino acids, and Martin's subsequent 1952 work (with Anthony James) extending the same partition principle to a gas mobile phase — inventing gas-liquid chromatography — together established that the mobile-phase axis (gas versus liquid) and the mechanism axis (adsorption versus partition) really are independent design choices, exactly the organizing insight this article builds on. Ion-exchange chromatography's quantitative mass-action treatment traces to work on synthetic resins during and after the Manhattan Project, where separating chemically similar rare-earth and actinide ions by charge-driven displacement chromatography was an urgent practical need; size-exclusion chromatography emerged later still, formalized by J. Porath and P. Flodin's 1959 development of dextran gel filtration media, extending the same distribution-equilibrium logic to a purely steric mechanism.