Thin-Layer and Paper Chromatography

Planar chromatography remains a first-line, low-cost screening tool wherever a fast go/no-go separation, rather than exhaustive quantitation, is the goal. Pharmaceutical quality control uses TLC identity tests specified in pharmacopeial monographs (USP, Ph. Eur.) to confirm that a drug substance co-migrates with a reference standard and to flag gross impurities before more expensive HPLC assays are run. Forensic chemistry uses TLC to screen seized substances for controlled drugs (opiates, cannabinoids, amphetamines) using standardized solvent systems and spray reagents (Marquis, Dragendorff) that give characteristic colors even before mass-spectral confirmation. Clinical toxicology historically used TLC "tox screens" for urine drug panels, and while immunoassay and LC-MS/MS have largely replaced routine use, TLC remains a low-resource-setting fallback and a teaching tool for the underlying separation logic.

Natural products and phytochemistry rely heavily on TLC fingerprinting — a single plate developed with a standardized solvent and visualized under UV/vis and with vanillin-sulfuric acid or anisaldehyde spray reagents gives a reproducible "chemical fingerprint" used to authenticate herbal medicines, detect adulteration, and guide fraction selection during natural-product isolation, well before any compound is purified enough for NMR. Lipid biochemistry uses TLC to separate phospholipid classes (phosphatidylcholine, phosphatidylethanolamine, sphingomyelin) by charge and headgroup polarity on silica developed with chloroform:methanol:water systems, visualized by charring or by molybdenum blue spray for phosphate-specific detection. Modern high-performance TLC (HPTLC), using finer, more uniform 5-μm silica particles and automated, volume-controlled spotting (nanoliter precision), pushes efficiency and reproducibility close to that of packed columns while retaining the parallel, multi-sample-per-plate throughput that makes planar chromatography uniquely fast for screening dozens of samples side by side. Coupled with slit-scanning densitometers reading absorbance or fluorescence directly off the plate, HPTLC gives quantitative results with relative standard deviations of 2–5%, feeding directly into the broader quantitative methodology developed in Qualitative and Quantitative Chromatographic Analysis. Instrument manufacturers now also couple TLC plates directly to mass spectrometers (TLC-MS interfaces, desorption electrospray ionization) so that a spot located by simple visual inspection can be identified structurally without ever being physically extracted from the plate — a direct, high-tech descendant of Tswett's original colored bands on calcium carbonate.

Ask the tutor: How do I choose a mobile-phase solvent system to hit a target Rf for a specific compound class on silica?

A TLC plate is a rigid backing — glass, aluminum foil, or plastic — coated with a thin layer (typically 100–250 μm) of finely divided adsorbent, most commonly silica gel (SiO₂·xH₂O) bonded with a small amount of calcium sulfate binder, or alumina (Al₂O₃), or a bonded reversed-phase silica (C18) for less polar separations. Paper chromatography uses cellulose fibers directly; the "stationary phase" here is largely the water adsorbed onto and within the cellulose matrix, so PC is fundamentally a partition technique even though it looks like adsorption chromatography at a glance. A sample is dissolved in a volatile solvent and applied as a small, concentrated spot (or narrow band, for preparative work) roughly 1.5–2 cm from the bottom edge of the plate — never touching the solvent, since the origin must stay above the liquid level in the developing tank. The plate is then placed in a closed chamber containing a shallow pool of developing solvent (the eluent), with its lower edge just submerged below the origin. Capillary action draws solvent upward through the porous layer; this is ascending development, the most common geometry. As the solvent front advances, it carries dissolved analyte molecules with it, but each molecule spends part of its time adsorbed onto (or partitioned into) the stationary layer and part of its time dissolved in the moving solvent. Molecules that adsorb strongly or partition poorly into the mobile solvent lag near the origin; molecules that adsorb weakly or dissolve preferentially in the mobile solvent are carried close to the solvent front. The chamber must be saturated with solvent vapor before and during development — otherwise evaporation from the plate surface creates a nonuniform, edge-effect-distorted front, and $R_f$ values become irreproducible. This is usually accomplished by lining the tank with filter paper wetted in solvent and allowing 15–30 minutes of equilibration before inserting the plate. Development is stopped when the solvent front has travelled a convenient distance (commonly 10–15 cm), the plate is removed, the front position is immediately marked in pencil (it fades as solvent evaporates), and the plate is dried. Spots are then located: colored compounds are visible directly; many organics fluoresce or quench fluorescence under 254 nm or 366 nm UV light against a fluorescent indicator embedded in the layer; and colorless, non-fluorescent compounds require a chemical visualization step — iodine vapor (forms reversible charge-transfer complexes with many organics), sulfuric acid char, ninhydrin for amines and amino acids, or a reagent targeted to a specific functional group.

The retardation factor is defined purely geometrically:

where $d_i$ is the distance travelled by the center of spot $i$ from the origin and $d_s$ is the distance travelled by the solvent front, both measured along the direction of development. Because $d_i \le d_s$ always, $R_f$ is dimensionless and bounded, $0 \le R_f \le 1$. A compound that never leaves the origin has $R_f = 0$; a compound that moves exactly with the solvent front (essentially unretained) has $R_f = 1$. The molecular picture behind this ratio is a dynamic equilibrium distribution, exactly as in column chromatography. At any instant, a fraction of analyte molecules sit in the stationary phase (adsorbed on silica, or dissolved in the stationary water layer of paper) and a fraction are dissolved in the mobile solvent occupying the pores of the layer. Only molecules in the mobile phase move; molecules in the stationary phase are momentarily immobile. The time-averaged fraction of molecules in the mobile phase determines how fast the spot's centroid advances relative to the solvent front — this is the same physical content as the retention factor $k$ used for columns, just expressed as a position ratio instead of a time ratio.

The central relationships connect the empirical, directly measured $R_f$ to the equilibrium distribution coefficient and to the geometric phase ratio of the layer.

where: $d_i$ = distance from origin to the center of spot $i$; $d_s$ = distance from origin to the solvent front, both measured at the moment development is stopped. Reading: $R_f$ is a position ratio at a fixed instant, not a rate — it is only meaningful relative to a specific, recorded solvent-front distance and development time.

where: $K = C_s/C_m$ is the equilibrium distribution (partition or adsorption) coefficient of the analyte between stationary phase concentration $C_s$ and mobile phase concentration $C_m$; $\beta = V_s/V_m$ is the phase ratio, the volume of stationary phase available per unit volume of mobile phase in the layer. Reading: as $K$ or $\beta$ grows (stronger retention, or more stationary phase per unit mobile-phase volume), $R_f$ shrinks toward 0; as $K\to 0$ (no retention), $R_f\to 1$. A second essential relationship connects $R_f$ to the retention factor $k$ used throughout column chromatography (from Retention Parameters and the Chromatogram):

where: $k$ is the same capacity/retention factor defined for a column, $k = (t_R - t_M)/t_M$; here it is recovered purely from spot geometry. Reading: $R_f$ and $k$ carry identical physical information — an $R_f$ of 0.50 corresponds to $k=1$ (equal time in each phase), while $R_f = 0.20$ corresponds to $k=4$ (four times as much time/mass in the stationary phase as in the mobile phase). Finally, relative migration between two spots is captured by a selectivity-like ratio, useful for predicting whether two compounds will resolve:

where: subscripts 1 and 2 denote the more- and less-retained spots respectively. Reading: two spots with the same $R_f$ cannot be separated regardless of plate length; the practical separating power of TLC is bounded by how large $\Delta R_f$ can be made through solvent choice, since (unlike a column) the plate length is fixed and short.

Example 1 — Foundational: computing Rf and k from a densitometer trace. A plate is developed with hexane:ethyl acetate (7:3) over $d_s = 9.40\ \text{cm}$ (measured from the pencil-marked origin to the solvent front). Three components of an over-the-counter analgesic tablet extract give spot centers at $d_1 = 1.60\ \text{cm}$ (caffeine), $d_2 = 4.42\ \text{cm}$ (acetaminophen), $d_3 = 7.99\ \text{cm}$ (aspirin), each measured to the nearest 0.01 cm with a millimeter ruler under UV light (254 nm, fluorescence-quenching layer). Compute each $R_f$:

Each is reported to three significant figures, consistent with the precision of the ruler (±0.01 cm on both $d_i$ and $d_s$ propagates to roughly ±0.002–0.003 in $R_f$ by standard error propagation: $\sigma_{R_f}/R_f \approx \sqrt{(\sigma_{d_i}/d_i)^2 + (\sigma_{d_s}/d_s)^2}$, which for spot 1 gives $\sigma_{R_f} \approx 0.170\sqrt{(0.01/1.60)^2+(0.01/9.40)^2} \approx 0.170 \times 0.0064 \approx 0.001$). Convert to retention factors using $k = (1-R_f)/R_f$:

Caffeine ($k_1 = 4.88$) is by far the most strongly retained — expected, since it is the most polar of the three and interacts most strongly with residual silanol groups on the silica. Aspirin, being relatively nonpolar in this solvent system, moves almost with the front ($k_3 = 0.176$). These $k$ values, though obtained from a five-minute plate rather than a 20-minute HPLC run, describe exactly the same equilibrium distribution physics.

Example 2 — Applied: two-dimensional paper chromatography of an amino acid mixture. A biochemistry student separates a five-component amino acid hydrolysate on a 20 cm × 20 cm cellulose sheet using two-dimensional development: first with phenol:water (75:25) in one direction ($d_{s,1} = 16.5\ \text{cm}$), then, after 90° rotation and thorough drying to remove phenol, with butanol:acetic acid:water (4:1:5, upper organic phase) in the perpendicular direction ($d_{s,2} = 14.8\ \text{cm}$). Spots are visualized with ninhydrin spray (0.2% in acetone, heated at 100 °C for 5 min), which reacts with primary amines to form the deep purple Ruhemann's purple chromophore. Two amino acids, glycine and alanine, are known to co-migrate ($R_f \approx 0.30$ in both solvents individually) when run one-dimensionally, because their $\Delta R_f$ in either single solvent is smaller than the plate's practical resolving increment. In the first dimension (phenol:water), glycine and alanine indeed overlap at $d_1 = 4.95\ \text{cm}$, giving

After rotation, development in butanol:acetic acid:water separates them because their relative polarities interact differently with the acidic, more hydrophobic second solvent: glycine (smaller, more polar side chain — in fact no side chain) moves $d_{2,\text{Gly}} = 3.55\ \text{cm}$, while alanine (methyl side chain, slightly more hydrophobic) moves further, $d_{2,\text{Ala}} = 5.77\ \text{cm}$:

The two-dimensional map now separates the spots by

corresponding to a spot-center separation on paper of $0.150 \times 14.8\ \text{cm} = 2.22\ \text{cm}$ — comfortably resolvable against typical amino acid spot diameters of 0.5–1.0 cm after ninhydrin development. This illustrates the central strategic tool of planar chromatography: when one solvent system cannot resolve two analytes because their $\Delta R_f$ is too small, a chemically orthogonal second solvent, applied in the perpendicular direction, frequently can — this is the entire logic of 2-D TLC/PC, later generalized into 2-D column methods.

Example 3 — Exam-level: optimizing solvent strength to hit a target $\Delta R_f$ and predicting resolution. A quality-control chemist must separate two structurally similar steroids, A and B, on silica TLC. In a nonpolar solvent (100% toluene), both run near the front: $R_{f,A} = 0.92$, $R_{f,B} = 0.95$ — essentially unresolved, since $\Delta R_f = 0.03$. Adding a polar modifier (ethyl acetate) to toluene slows both spots by increasing $K$ (stronger adsorption competition is reduced less selectively for A than B, since A has one more hydrogen-bond donor than B). Empirically, for this modifier system, retention follows an approximately linear relationship between $\log k$ and volume fraction of ethyl acetate, $\phi$, over the workable range:

(the two lines have the same slope — a common simplification for chemically similar analytes with the same functional-group interaction mechanism but different affinities — and are separated by a constant intercept difference of 0.25 in $\log k$, i.e., a constant selectivity $\alpha = k_A/k_B = 10^{0.25} = 1.78$ at all $\phi$.) The chemist wants both spots to fall in the practically optimal $R_f$ window, $0.2 \le R_f \le 0.5$ (spots too close to the origin or the front are hard to measure precisely and more prone to tailing/streaking artifacts). Target the less-retained compound, B, at $R_f = 0.35$, near the middle of this window.

Step 1. Convert target $R_f$ to target $k_B$:

Step 2. Solve for $\phi$ from the B equation:

So a mobile phase of roughly 27% ethyl acetate in toluene ($\phi = 0.268$) should place compound B at $R_f \approx 0.35$.

Step 3. Predict $k_A$ and $R_{f,A}$ at this same $\phi$, using the constant selectivity relationship, $\log k_A = \log k_B + 0.25$:

Step 4. Check separation quality. The predicted spot positions are $R_{f,A} = 0.232$ and $R_{f,B} = 0.350$ over a development distance of, say, $d_s = 10.0\ \text{cm}$, giving spot centers at $2.32\ \text{cm}$ and $3.50\ \text{cm}$ — a center-to-center separation of $1.18\ \text{cm}$. If typical spot diameters (after development, accounting for diffusion broadening) are about $0.6$–$0.8\ \text{cm}$, this separation is adequate for baseline resolution by eye or densitometer, though tight enough that chamber saturation and constant loading (to avoid streaking that would merge the spot edges) become essential.

Step 5 — limiting-case sanity check. If instead the chemist had left $\phi = 0$ (pure toluene), the model predicts $\log k_B = 1.10 \Rightarrow k_B = 12.6 \Rightarrow R_{f,B} = 1/13.6 = 0.0735$ — but the problem stated that the observed $R_{f,B}$ at $\phi=0$ was 0.95, not 0.07. This discrepancy is expected and instructive: the linear $\log k$ vs. $\phi$ model is a local approximation valid near the operating region used to fit it (here, near $\phi \approx 0.2$–$0.3$), and it should never be extrapolated to $\phi=0$ without new calibration data — a caution any real optimization must respect, since solvent-strength/retention relationships in adsorption TLC are only piecewise, not globally, linear.

Planar chromatography's modern form owes its greatest single push to Mikhail Tswett, the Russian-Italian botanist who in 1901–1903 first separated plant pigments by passing petroleum-ether extracts through columns of powdered calcium carbonate, watching colored bands separate — coining the word "chromatography" (color-writing) for the resulting bands. Tswett's original technique was columnar, not planar, but the underlying physical insight — that differential adsorption from a flowing solvent could separate closely related compounds into visibly distinct zones — is identical to what happens on a modern silica plate. Tswett worked largely in obscurity, his findings disputed and even dismissed by contemporaries including Richard Willstätter (who initially could not reproduce clean separations and publicly doubted the method), and it took until the 1930s, when Edgar Lederer and others revived and refined column adsorption chromatography for carotenoid separations, for the technique's power to be widely recognized. The planar (thin-layer) format as routinely practiced today was substantially developed and popularized by Egon Stahl in the late 1950s, who standardized layer thickness, activation temperature, spotting technique, and solvent systems into a reproducible protocol — his 1969 handbook Thin-Layer Chromatography became the reference manual that turned TLC from an idiosyncratic craft into a routine analytical method used in every organic chemistry teaching laboratory worldwide.