The Chemistry Behind Chromatography
Chromatography looks like a laboratory skill and is really an equilibrium problem. Every separation, from a TLC plate in a Class 12 practical to a two-metre protein column, rests on one idea: each component distributes itself between a moving phase and a stationary phase, and the components that spend more time stuck move more slowly. Everything else — Rf, retention time, plate number, resolution — is bookkeeping on top of that single equilibrium. This article does the bookkeeping properly, with the arithmetic worked out, and then states clearly what the technique cannot do.
The formulas you already know
K = cstationary ÷ cmobile (partition coefficient)
k = (tR − tM) ÷ tM (retention factor)
α = k₂ ÷ k₁ (selectivity) N = 16 (tR/w)² (plate number)
Rs = 2(tR2 − tR1) ÷ (w₁ + w₂) (resolution)
What each term means
| Term | Meaning | Unit |
|---|---|---|
| K | How strongly a component prefers the stationary phase at equilibrium | — |
| tM | Dead time — time for an unretained species to pass through | min |
| tR | Retention time of the component | min |
| k | Retention factor — how many extra dead times the component takes | — |
| α | Selectivity — how differently the phase treats two components | — |
| N | Plate number — a measure of column efficiency (peak sharpness) | — |
| H | Plate height, L/N — the length needed for one theoretical equilibration | µm or cm |
| Rs | Resolution; Rs ≥ 1.5 is baseline separation | — |
k and K are linked by the geometry of the column: k = K × (Vstationary ÷ Vmobile). K is chemistry — it depends only on the two phases, the analyte and the temperature. k is what your instrument actually reports, and it also depends on how much stationary phase is packed inside.
Partition or adsorption — the two mechanisms
Partition means the analyte genuinely dissolves in a liquid stationary phase coated on or bonded to a solid support. Separation follows relative solubility, driven by the ordinary interactions: hydrogen bonding, dipole–dipole, dispersion forces. Adsorption means the analyte sticks to the surface of a solid such as silica gel or alumina, held by its surface hydroxyl groups. Silica is polar, so polar analytes are held hardest and move least — the basis of "normal phase". Bond a long hydrocarbon chain onto that silica and the whole scale inverts: non-polar analytes are now retained hardest, which is reverse phase, the most widely used mode in modern liquid chromatography. Ion exchange (charge), size exclusion (molecular size) and affinity (biological recognition) are further variants of the same distribution idea.
Worked example 1 — Rf from a TLC plate
The solvent front travels 8.0 cm from the baseline. Two spots are seen at 3.6 cm and 5.4 cm.
Rf(A) = 3.6 ÷ 8.0 = 0.45
Rf(B) = 5.4 ÷ 8.0 = 0.68
Spot B has the higher Rf, so it spends less time on the stationary phase. On a silica plate that means B is the less polar of the two.
The planar equivalent of the retention factor follows directly:
k = (1 − Rf) ÷ Rf = (1 − 0.45) ÷ 0.45 = 0.55 ÷ 0.45 = 1.22 for spot A.
Rf is always between 0 and 1. A spot at the baseline (Rf = 0) is too strongly held; a spot on the front (Rf = 1) is not held at all. Both mean the solvent system is wrong — good separations sit roughly in the 0.2–0.8 band.
Worked example 2 — retention, efficiency and resolution
A column gives a dead time tM = 1.20 min. Peak 1: tR = 6.40 min, base width w = 0.44 min. Peak 2: tR = 7.20 min, w = 0.50 min.
Retention factors:
k₁ = (6.40 − 1.20) ÷ 1.20 = 5.20 ÷ 1.20 = 4.33
k₂ = (7.20 − 1.20) ÷ 1.20 = 6.00 ÷ 1.20 = 5.00
Selectivity: α = 5.00 ÷ 4.33 = 1.154 — above 1, so the two really are treated differently by the phase.
Efficiency (peak 2):
N = 16 × (7.20 ÷ 0.50)² = 16 × (14.4)² = 16 × 207.36 = 3318 plates
On a 25 cm column, H = L ÷ N = 25 ÷ 3318 = 0.00753 cm = 75 µm.
Resolution:
Rs = 2 × (7.20 − 6.40) ÷ (0.44 + 0.50) = 1.60 ÷ 0.94 = 1.70
Above 1.5, so the peaks are baseline-separated and can be integrated independently.
Cross-check by a second route. The master resolution equation splits Rs into its three controllable parts:
√3318 = 57.6, so √N/4 = 14.4
(α − 1)/α = 0.154 ÷ 1.154 = 0.1333
k₂/(1 + k₂) = 5.00 ÷ 6.00 = 0.8333
Rs = 14.4 × 0.1333 × 0.8333 = 1.60
Not quite the 1.70 above — and the difference is instructive, not an error. The master equation assumes both peaks have the same width, and here peak 1 is narrower (0.44 min against 0.50 min). Substituting w = 0.50 for both in the direct formula gives 2 × 0.80 ÷ 1.00 = 1.60 exactly. Whenever two valid routes disagree, look for the assumption one of them made.
The master equation is also the practical guide to fixing a poor separation. Because Rs depends on √N, doubling the column length improves resolution only by about 1.41 times while doubling the run time. Changing the chemistry to improve α is far more powerful, and raising a very low k helps a great deal — though beyond k ≈ 10 the k/(1+k) term is nearly 1 and longer retention buys almost nothing.
Where this is actually used
Chromatography is the routine separation and quantification tool of essentially every analytical laboratory. Pharmaceutical quality control uses liquid chromatography for the assay of an active ingredient and for impurities and degradation products. Food testing screens for pesticide residues and adulterants; air and water laboratories separate volatile organics by gas chromatography; petroleum laboratories resolve hydrocarbon mixtures of enormous complexity; forensic and clinical toxicology rest on it entirely. In biology, preparative ion-exchange, size-exclusion and affinity columns are how a protein is purified in the first place.
In practice the separation is almost always coupled to a detector that provides identity as well as amount — most commonly a mass spectrometer, giving GC-MS and LC-MS. That combination exists precisely because of the limitation in the next section.
The honest limits
- Rf is not a physical constant. It depends on the plate batch, the solvent composition, chamber saturation, temperature, humidity, how much material was spotted and how far the front was allowed to run. An Rf from a textbook cannot identify your spot. The only sound use is a reference standard run on the same plate, side by side with the unknown.
- Retention time identifies nothing on its own. Many compounds share a retention time on any given column. A matching tR is consistent with an identity; it does not establish one. Confirmation needs a mass spectrum, a second column of different selectivity, or an independent spectroscopic method.
- Co-elution hides components. One symmetrical peak can be two substances, so a "clean" chromatogram is never proof of a pure sample.
- Peak area is not amount until you calibrate. Detector response differs from compound to compound, so area percent is not composition percent. Quantitative work needs standards, and usually an internal standard.
- Bands broaden as they travel. Efficiency is limited by eddy diffusion, longitudinal diffusion and slow mass transfer between phases — the three terms of the van Deemter relation, H = A + B/u + Cu. Because of the middle term there is an optimum flow rate: running faster than it costs resolution, and so does running slower.
- Overloading destroys the peak shape. Inject too much and the distribution stops being linear, peaks front or tail, and both N and Rs collapse. Tailing on silica is often caused by residual surface hydroxyl groups interacting with basic analytes — a chemistry problem, not a hardware fault.
- Not everything can be run. Gas chromatography needs the analyte to be volatile and thermally stable; involatile or fragile compounds must be derivatised or moved to a liquid method.
- Measure tM, do not guess it. Every k, α and the master equation depend on the dead time. An assumed tM makes all of them wrong together, in a way the numbers will not reveal.
Why this matters for JAM, GATE, NET and CUET-PG
| Exam area | What is typically asked |
|---|---|
| Analytical chemistry | Rf, k, α, N and Rs calculations from given data |
| Separation techniques | Normal vs reverse phase; which component elutes first and why |
| Column efficiency | Van Deemter terms and the optimum flow rate |
| Instrumental analysis | GC vs HPLC suitability; detectors; hyphenated techniques |
| Practical/viva | Choice of solvent system, spotting, chamber saturation, visualisation |
The arithmetic above is where marks are lost, not the theory. Ratios, square roots, plate numbers and calibration slopes are all quick to check in the calculator suite — it also holds the physical-chemistry tools you will need alongside a separation, from Beer–Lambert to linear regression.
Open the ABC Chemistry Calculator Suite →Preparing for IIT-JAM, GATE, CSIR-NET or CUET-PG chemistry? ABC Chemistry runs dedicated competitive-exam batches at the coaching centre and online across India — details at abcchemistry.in.