Chromatographic Techniques: TLC, HPLC, HPTLC and GLC

Chromatography works because every analyte partitions between two phases with its own equilibrium constant. Push the mobile phase across the stationary phase and the analyte spends part of its time absorbed onto the stationary phase (stationary, going nowhere) and part of its time dissolved in the mobile phase (moving with the flow). If the equilibrium favours the stationary phase the analyte moves slowly; if it favours the mobile phase the analyte moves quickly. Two compounds with even slightly different partition coefficients will separate over enough column length.

The only thing that differs across TLC, HPLC, HPTLC and GC is the geometry and the speed. TLC and HPTLC develop a planar plate where the eye reads position. HPLC and GC push the mobile phase through a long thin column where a detector at the far end reads time. The position on a TLC plate and the time on a GC chromatogram are the same physical quantity expressed differently: a measurement of how slowly the analyte was retained by the stationary phase.

Forensic identification uses three numbers from this physics. Rf on a TLC plate (distance moved by the spot divided by distance moved by the solvent front). Retention time tR on an HPLC or GC chromatogram. Resolution Rs between adjacent peaks, with 1.5 as the conventional baseline-separation threshold. A forensic match means the unknown produces the same Rf or the same tR as a reference standard on the same instrument with the same method, ideally backed by a second confirmatory technique on a different physical principle.

TLC is the simplest and oldest of the four. A glass or aluminium plate is coated with a thin layer of silica gel 60 F254 (the F254 means the silica contains a fluorescent indicator that lights up under 254 nm UV; analytes that absorb in that range appear as dark spots against a bright green background). Standard plate sizes are 20 by 20 centimetres for full screening and 5 by 10 centimetres for spot checks. The sample extract is applied as a small dot one centimetre from the bottom edge, the plate is placed in a glass tank containing a few millilitres of solvent, the solvent climbs the plate by capillary action and the analytes ride along at different speeds.

The choice of solvent system depends on the chemistry of the suspected analyte. Indian toxicology manuals (Modi's Medical Jurisprudence and Toxicology, the BPRD forensic handbook and the CFSL SOP set) standardise on four mobile phases that between them cover most drug classes.

Once the plate has developed, the analyst dries it and visualises the spots. UV at 254 nm is the first look, picking up any analyte that quenches the fluorescent silica indicator. Spray visualisation follows, chosen for the suspected class. Marquis reagent (formaldehyde plus concentrated sulphuric acid) gives a purple colour with opioids and an orange colour with amphetamines. Mayer's reagent (potassium mercuric iodide) precipitates alkaloids as a cream-white spot. Dragendorff's reagent (potassium bismuth iodide) gives an orange-brown spot with most nitrogenous bases. Ninhydrin develops amino acids and amphetamines as purple-pink. Iodine vapour in a sealed tank gives a generic brown spot for any unsaturated organic compound.

Rf comparison with co-spotted reference standards is the working logic. A diazepam reference run on the same plate at the same time as the unknown gives an internal Rf yardstick that cancels out plate-to-plate variation in solvent strength, humidity and temperature. The unknown spot at Rf 0.42 next to a diazepam reference also at Rf 0.42 is a strong screening result. It is not, by itself, a confirmation; the limitation is sensitivity (approximately five micrograms of analyte per spot is the lower threshold of reliable visualisation) and specificity (several drugs share Rf values in any given system, which is why two solvent systems should always be run in parallel).

HPLC moved the planar idea into a closed steel column under pressure. The stationary phase is a finely packed bed of silica beads, chemically modified so the surface is non-polar. Reverse-phase C18 columns (a layer of octadecyl chains bonded to the silica) are the default for forensic drug analysis. The typical analytical column is 250 millimetres long and 4.6 millimetres in internal diameter, packed with five micrometre particles. UHPLC variants run shorter and narrower (100 by 2.1 millimetres) with 1.8 micrometre particles, trading higher backpressure for faster runs and better resolution.

The mobile phase is a mixture of water and an organic modifier (methanol, acetonitrile) often pulsed across a gradient so the solvent strength rises during the run and pulls progressively more retained analytes off the column. A small concentration of formic acid or an acetate buffer keeps the pH stable so weakly basic drugs do not develop tailing peaks. A high pressure pump (Shimadzu Prominence LC-2030, Agilent 1100 or 1260, Waters Alliance are the three platforms most state SFSLs and medical college toxicology departments actually own) drives the mobile phase at one to two millilitres per minute.

What separates HPLC from GC is the detector zoo at the far end. Each detector picks up a different physical signature and the choice depends on the analyte.

The standing Indian SFSL workflow on HPLC-DAD reads roughly like this. Barbiturates in viscera or post-mortem blood are quantitated on an isocratic phosphate-buffer methanol method at 220 nm. Paracetamol overdose cases run on a 250 nm channel. Alprazolam in drug-facilitated crime samples is screened on a reverse-phase gradient at 230 nm and confirmed against a deuterated internal standard for high-value cases. Antidepressants, antihistamines and beta-blockers each have a worked-up SOP at CFSL and the better state SFSLs.

UHPLC has entered Indian forensic practice over the last decade. Shorter columns and smaller particles cut a 30-minute LC run to under eight minutes, which matters when a state SFSL clears two hundred backlog cases a quarter. The trade-off is higher pressure (over 800 bar), shorter column life and a more demanding sample preparation step.

HPTLC is not a different technique from TLC so much as TLC done with engineering and automation. The silica particle size drops from ten to fifteen micrometres in regular TLC to three to five micrometres in HPTLC, which sharpens spots and roughly doubles resolution. The sample is applied as a narrow band (not a dot) by an automated band applicator like the Camag Linomat 5, which sprays a precisely metered volume across a defined band width using a syringe and a nitrogen stream. The plate is developed in an automated developing chamber (Camag ADC2) that controls humidity, saturation and solvent front travel. After development, the plate goes onto a densitometric scanner (Camag TLC Scanner 4) that reads absorbance or fluorescence at chosen wavelengths across each band, producing a chromatogram that looks much like an HPLC trace but with band position as the x-axis.

The result is a planar technique that is quantitative (peak area on the densitogram correlates linearly with analyte mass over a calibration range), reproducible (because everything from application to scanning is instrumental rather than manual) and capable of running up to seventy samples in parallel on a single 20 by 10 centimetre plate. That throughput is why HPTLC is the workhorse of Indian institutions that screen herbal and food matrices.

In forensic and regulatory practice HPTLC is concentrated at NIPER Mohali for pharmaceutical fingerprinting and adulteration cases, NIN Hyderabad for nutritional and food safety profiling, CCRAS for ayurvedic and herbal fingerprinting under the Drugs and Cosmetics Rules, and FSSAI-empanelled laboratories for food adulterant screening (synthetic dyes in chilli powder, melamine in milk, mineral oil in edible oils).

The relevance to a forensic toxicology lab is twofold. Suspected herbal poisoning cases (aconite from Aconitum napellus, Cerbera odollam from yellow oleander, abrin from Abrus precatorius) often need a fingerprint comparison rather than a single-analyte assay, and HPTLC is faster and cheaper than HPLC-MS for that job. Dye and pigment forensics in ink, paint and textile fibre cases also lean heavily on HPTLC because dyes are non-volatile and separate cleanly on silica.

Gas-liquid chromatography pushes a vaporised sample through a long thin capillary column with a flowing inert gas. The forensic standard column is fused silica, 30 metres long, 0.25 millimetres in internal diameter, with a 0.25 micrometre liquid film coated on the inner wall. The stationary phase varies with target chemistry: DB-1 (100 percent polydimethylsiloxane, non-polar, broad screen), DB-5 (5 percent phenyl, the universal forensic phase, used for drugs and pesticides) and DB-WAX (polyethylene glycol, polar, ideal for alcohols and short-chain volatiles).

Helium is the traditional carrier gas in Indian labs. Hydrogen is the cheaper and faster alternative and is now common at CFSL Hyderabad and several refurbished state SFSLs, with appropriate cylinder safety controls.

The detector dictates what the GC actually does for a case.

Two GC sample-introduction modes carry most forensic volatile work. Headspace GC takes a sealed glass vial containing the biological matrix (blood, urine, gastric contents) and heats it to a fixed temperature. Volatile analytes partition into the vapour above the liquid, and a sample of that headspace gas is injected onto the column. Ethanol in blood is the canonical headspace assay, run on every drink-drive case routed through an SFSL. The same workflow handles methanol (Bihar 2016, Gujarat 2022, Tamil Nadu 2024 hooch tragedies), chloroform (the classical criminal stupefying agent), petrol and other organic solvents.

For non-volatile drugs, derivatisation makes the analyte sufficiently volatile for GC analysis. Derivatisation converts polar hydroxyl, carboxyl or amino groups into less polar silyl or methyl groups so the molecule survives the GC inlet. Silylation with BSTFA (N,O-bis(trimethylsilyl)trifluoroacetamide) is the workhorse for hydroxylated drugs (cannabinoids, opioid metabolites, benzodiazepine metabolites). Methylation with diazomethane covers carboxylic acid drugs (some NSAIDs, valproate). Derivatisation is the bottleneck of GC sample prep and is one reason LC-MS/MS has gradually eaten into GC casework at tier-one Indian labs.

GC-MS remains the courtroom gold standard for drug and poison identification in Indian forensic practice. The combination of retention time on a defined column, electron-impact mass spectrum and library match (NIST and Wiley libraries) is the evidentiary trifecta. AIIMS forensic toxicology, NIMHANS Bangalore, CFSL Chandigarh, CFSL Hyderabad and FSL Madhuban (Haryana) all run GC-MS as their primary confirmatory platform. LC-MS/MS at NDTL (the WADA-accredited National Dope Testing Laboratory), AIIMS and a handful of metro labs handles the cases GC-MS cannot reach: thermally labile drugs, polar metabolites and ultra-trace post-mortem casework.

The choice between TLC, HPLC, HPTLC and GC in Indian forensic practice is rarely free. It is dictated by what the lab actually owns, by the chemistry of the suspected analyte and by the evidentiary standard the prosecution will need to meet.

1. Start with the suspected class
   Volatile poisons (ethanol, methanol, chloroform, organic solvents, petrol residues from arson scenes) go to headspace GC. Pesticides and explosives go to GC with ECD or NPD. Thermally fragile drugs (cannabinoids, opioids, benzodiazepines, barbiturates) go to HPLC-UV or HPLC-MS. Herbal and dye fingerprints go to HPTLC.
2. Run a TLC screen if equipment is limited
   At district SFSLs without GC-MS or LC-MS, TLC in two solvent systems (System A and one drug-specific system) plus targeted spray reagents will still resolve most casework to a workable screening report, often within the same day.
3. Quantitate with HPLC-DAD or GC-FID
   Most state SFSLs are equipped with HPLC-DAD and GC-FID and can quantitate the confirmed analyte against a calibration curve with a deuterated or structurally analogous internal standard.
4. Confirm at CFSL, AIIMS or NDTL
   Cases that will go to trial, cases with disputed chain of custody, and cases of trace post-mortem toxicology should be routed to a GC-MS or LC-MS/MS facility. CFSL Chandigarh, CFSL Hyderabad, AIIMS forensic toxicology and NDTL Delhi are the standard referral nodes.
5. Report retention parameters and the second technique
   The forensic report should record Rf or retention time, the column and mobile phase used, the reference standard run alongside the sample, the detector response and the confirmatory technique with its spectral or fragmentation match. A report that lists only a single technique without a second-principle confirmation is vulnerable on cross-examination.

Equipment distribution matters when planning casework. Every state SFSL has at least one GC-FID and one HPLC-DAD as a working minimum. GC-MS is concentrated at CFSL Chandigarh, CFSL Hyderabad, FSL Madhuban, AIIMS and NIMHANS, with a slow rollout to tier-one state labs. LC-MS/MS is rarer still, anchored at NDTL Delhi, AIIMS and the CFSL trio. HPTLC sits outside the formal SFSL network at NIPER Mohali, NIN Hyderabad and CCRAS-funded centres. TLC is universal because the consumables cost almost nothing.

A forensic chromatogram has to survive cross-examination. Three classic failure modes account for most of the rejected runs in Indian SFSL audits.

Tailing peaks are the commonest complaint, especially with basic drugs (amphetamines, ephedrine, codeine, alprazolam). The cause is almost always silanol activity on the column packing: free Si-OH groups on the silica surface form hydrogen bonds with basic nitrogen atoms in the analyte, delaying its release. The fixes are predictable. Add an ammonia or triethylamine modifier (about 0.1 percent) to the HPLC mobile phase to mask the silanols. Switch to a deactivated or end-capped C18 column. On GC, use a base-deactivated capillary phase. If the tail persists, the column is degraded and needs replacement; a long, drifting tail with each successive injection on the same column is the telltale.

Poor resolution between two adjacent peaks (Rs below 1.5) is the second failure mode. The fixes are method-dependent. On HPLC, run a longer gradient with a shallower slope so the two analytes spend more time at their respective elution strengths. On GC, raise the initial oven temperature more slowly and increase the column length or split ratio. On TLC, switch to a less polar solvent system so the spots travel less far and have more room to separate. Resolution problems that survive method adjustment usually mean the column is approaching end-of-life.

Ghost peaks are extra peaks with no analyte in the injection. The causes split between carryover (the previous sample is still bleeding off the injector or column) and contamination (dirty inlet liner, degraded septum, impure mobile phase). Fixes are blanks between samples, a fresh inlet liner and septum every fifty injections on GC, and HPLC-grade solvents only. Ghost peaks at the retention time of the target are dangerous because they produce false positives; an unexplained peak in a procedural blank is a reason to invalidate the whole batch.

The four techniques together cover every routine forensic toxicology workflow in India, from a TLC screen at a district SFSL to GC-MS confirmation at CFSL Chandigarh and LC-MS/MS quantitation at AIIMS. Technique selection follows the analyte chemistry; detector selection follows the analyte's physical signature; and the forensic report must document both in terms a sessions court can evaluate.