Infrared Spectroscopy: FTIR and ATR-FTIR

A molecular bond is a spring. It has a natural frequency of vibration set by the masses at each end and the stiffness of the bond. An IR photon whose energy matches that natural frequency is absorbed and the bond vibrates harder. Absorbance at the matched wavenumber is what the spectrum draws.

There is one selection rule worth committing to memory. A vibration is IR active only if it changes the dipole moment of the molecule. Homonuclear diatomics like N2, O2 and H2 have no permanent dipole and no change in dipole on stretching, so they are completely IR invisible. Heteronuclear bonds (C=O, O-H, N-H, C-Cl) are strongly IR active. This is why ambient air does not interfere with the spectrum despite being mostly nitrogen and oxygen, and why CO2 at 2350 cm-1 and water vapour at 3700 and 1600 cm-1 do show up if the optical path is not purged.

The infrared spectrum splits into three regions and the working bench cares about one of them.

The wavenumber convention is worth being comfortable with. cm-1 is the reciprocal of wavelength expressed in centimetres. Higher wavenumber is higher photon energy and shorter wavelength. A C-H stretch at 3000 cm-1 is more energetic than a C-O stretch at 1100 cm-1. The axis runs right to left on a conventional spectrum, with the high wavenumber stretching modes on the left and the low wavenumber fingerprint region on the right.

A mid-IR spectrum reads in two halves. Above about 1500 cm-1 sit the group frequencies, where individual functional groups absorb at predictable wavenumbers regardless of the rest of the molecule. Below 1500 cm-1 sits the fingerprint region, where coupled and skeletal vibrations produce a pattern unique to the whole molecule.

The group frequency half is what the analyst eyeballs first. A broad band at 3200 to 3600 cm-1 means hydrogen-bonded O-H, almost always alcohol or carboxylic acid or water of crystallisation. A sharp band at 3300 to 3500 cm-1 with one or two prongs is N-H, primary or secondary amine. A strong absorption between 1680 and 1750 cm-1 is the carbonyl region, the single most diagnostic band in the whole spectrum, distinguishing aldehydes, ketones, esters, amides and acids.

The fingerprint region from 1500 to 400 cm-1 is where the molecular identity actually lives. Two compounds can share every group frequency above 1500 (paracetamol and a structurally related impurity often do) and still be distinguishable in the fingerprint pattern. A library search algorithm scores the match across the entire spectrum but weights the fingerprint region heavily, which is why a match score above 0.95 against a NIST or Sadtler library is the working confirmation threshold.

Older dispersive IR instruments used a monochromator with a grating that scanned wavelengths one at a time onto a slit. Acquisition was slow, signal-to-noise was poor at the low energy mid-IR end, and a single survey scan took several minutes. A Perkin Elmer 137 or 257 dispersive instrument could take several minutes for a single survey scan, the chart recorder crawling through 4000 to 400 cm-1.

FTIR replaced the monochromator with a Michelson interferometer. A beamsplitter divides the source beam into two arms, one with a fixed mirror and one with a moving mirror. The recombined beam carries every wavelength simultaneously, modulated at a different audio frequency depending on its wavenumber. The detector reads an interferogram (intensity versus mirror displacement) and a Fourier transform converts that into the conventional spectrum.

Two specific advantages drove the technology shift. The Felgett or multiplex advantage means every wavenumber is measured during every moment of the scan, so signal-to-noise improves with the square root of the integration time. The Jacquinot or throughput advantage means the optical path has no narrow slit, so much more light reaches the detector. Together these give a 20 to 100 fold improvement in scan time and signal-to-noise, and they are why FTIR has been the default since the late 1980s.

A typical bench FTIR scans 4000 to 400 cm-1 in 10 to 30 seconds at 4 cm-1 resolution. Co-adding 16 or 32 scans (still under a minute total) gives a clean spectrum on milligram quantities. Compared to dispersive instruments, this collapses a 30-minute scan into a casework workflow practical for high-volume throughput.

The classical solid mode is the KBr pellet. Potassium bromide is IR transparent across the entire mid-IR range, ductile enough to flow under pressure, and chemically inert with most analytes. About 1 to 2 mg of dry analyte is ground with roughly 200 mg of dry KBr in an agate mortar, transferred into a 13 mm stainless steel die, and pressed at 8 to 10 tonnes for two to three minutes. The disc that pops out is glass clear if the work was done well, milky if the powder was damp.

Nujol mull is the alternative for moisture sensitive solids. The analyte is ground with mineral oil into a paste, smeared between two NaCl or KBr plates and run as a suspension. The trade-off is that Nujol itself absorbs in the C-H stretch region (about 2900 cm-1) and the C-H bend region (1460 and 1377 cm-1), so those windows are blocked. For a chlorinated or oxidised analyte where Nujol is silent, the mull is fine; for a hydrocarbon, the mull is useless and a fluorocarbon oil (Fluorolube) is substituted.

Liquid mode runs neat liquids as a thin film between two NaCl or KBr plates with no spacer, or in a fixed pathlength solution cell of 0.025 to 1 mm built from KBr or CaF2 windows. Gas cells with 10 cm or longer optical paths handle vapours for atmospheric monitoring. None of these is used much in routine forensic chemistry.

What changed everything is ATR. A diamond, ZnSe or Ge crystal of refractive index well above the sample is mounted with the sample pressed against its top face. The IR beam hits the crystal-sample interface at greater than the critical angle and undergoes total internal reflection, but a small evanescent wave penetrates 0.5 to 2 µm into the sample and is attenuated at the wavenumbers the sample absorbs. The reflected beam, now carrying the IR spectrum of the surface layer, goes to the detector.

The ATR advantage is operational rather than spectroscopic. A sample placed on the diamond and pressed with a torque-limited anvil gives a usable spectrum in under a minute, with no sample destruction, no dilution, and no clean-up beyond wiping the crystal with isopropanol. Indian SFSL chemistry benches have standardised on diamond ATR for routine identification, with the KBr pellet retained for cases where the surface layer is not representative or a longer effective pathlength is needed.

Forensic chemistry casework is dominated by unknown solids requiring initial identification. IR answers that question faster than any other bench technique, which is why the same instrument appears across several different investigation streams.

Unknown white powder triage is the bread and butter. A constable walks in with a polythene packet seized from a roadside bag and the bench needs an answer before the investigation team commits to a narcotic versus non-narcotic line. ATR-FTIR distinguishes paracetamol from caffeine from sucrose from talc from ammonium nitrate from urea in seconds, and from methamphetamine, MDMA, cocaine and heroin against a controlled-substance library inside a minute. The result triages the sample to the toxicology bench for confirmation, the explosives bench for a different workup, or back to the IO with a non-suspicious finding.

Suicide-note tablets recovered at autopsy are an almost daily case at any tertiary medical college in India. A tablet with no markings, half dissolved in stomach contents, gets placed on the diamond and identified against a pharmaceutical library: aluminium phosphide, chloroquine, dapsone, paracetamol, organophosphorus tablet formulations. A confirmed identification in 30 seconds points the post-mortem chemistry workflow at the right specific assay (GFAAS for the metallic phosphide residue, GC-NPD for the organophosphate parent, HPLC for the hepatotoxin).

Counterfeit pharmaceuticals fall under CDSCO investigations and reach the FSL when state drug controllers seize a suspect batch. ATR-FTIR fingerprints the active pharmaceutical ingredient (API) against the bulk-drug reference, and any deviation in the fingerprint region flags adulteration, substitution or absence of the labelled API. NIPER Mohali maintains one of the better Indian bulk-drug ATR-FTIR libraries for exactly this work.

Paint chip layer-by-layer analysis is the classic hit-and-run case. A paint flake from a victim's clothing or a recovered vehicle is examined under a stereo microscope, the layers are separated mechanically (clear coat, base coat, primer, electrocoat, original factory layers), and each layer is mounted on a diamond ATR or on an IR microscope stage. The layer sequence and the IR fingerprint of each layer is matched against the reference paint flake from the suspect vehicle. A match across all layers in the same order is strong source attribution evidence, particularly for older multi-coat finishes.

Fibre identification distinguishes synthetic from natural and tells synthetics apart. Polyester (PET) shows a strong ester C=O at 1715 cm-1 and aromatic bands at 1410 and 720 cm-1. Nylon shows the secondary amide N-H at 3300 and amide I and II at 1640 and 1540 cm-1. Acrylic shows nitrile C≡N at 2240 cm-1. Cotton, viscose and other cellulosics show the broad O-H, C-H and ring-stretch pattern of cellulose. An IR microscope coupled to FTIR with about 10 µm spatial resolution handles single fibre cross sections without dissecting the fibre.

Plastic identification at recycling, smuggling and contamination cases distinguishes polyethylene (PE), polypropylene (PP), polyvinyl chloride (PVC), polystyrene (PS), PET and ABS in seconds. The wavenumber positions of the C-H stretches and the strong skeletal bands below 1500 cm-1 are diagnostic. A counterfeit pharmaceutical blister, a tampered food packaging, or a bag holding contraband is identified by polymer type before any further chemistry.

Explosive residue is a non-trivial application. TATP (triacetone triperoxide), RDX, HMX, PETN, ANFO (ammonium nitrate fuel oil) and urea nitrate each give characteristic IR spectra. ATR-FTIR on a swab extract or a recovered solid is used at CFSL Pune and the explosive units of state SFSLs as a presumptive identification. Confirmation rides on LC-MS/MS or ion chromatography for nitrate and nitrite, but the ATR-FTIR call is what triggers the confirmatory workflow.

Forged document examination uses IR to distinguish toner inks (carbon black with polymer binder) from liquid inks (dye in solvent), to compare the toner of a questioned page against a known printer source, and to identify the polymer of a security feature on a counterfeit currency note. An IR microscope examines the toner layer in situ on the document.

A modern FTIR ships with a software search engine and one or more spectral libraries. The NIST IR library, the Sadtler commercial libraries (now Bio-Rad KnowItAll), the Mainlib drug library and instrument-vendor in-house libraries between them cover most of the molecules a forensic bench will see. A typical search returns the top 10 hits ranked by a similarity score (Euclidean, first-derivative or correlation algorithm depending on software). A score above 0.95 is the working confirmatory threshold for routine identification, although a careful analyst still eyeballs the overlay rather than trusting the number alone.

Limitations matter. Aqueous samples are awkward because water absorbs strongly across most of the mid-IR range and overwhelms most analyte bands. Mixtures are the harder case, because overlapping bands from multiple components confuse both the eye and the search engine; the spectrum of a 50:50 paracetamol-caffeine tablet is not the sum of two clean spectra and a single library match is unreliable. Quantitation is possible by Beer-Lambert in the IR but is much less precise than UV-Vis quantitation, so the bench reaches for HPLC or UV-Vis when a number rather than an identity is needed.

The IR microscope couples a polarising or stereo microscope to the FTIR optical path with a small ATR objective or transmission stage. Spatial resolution is about 10 µm at the diffraction limit, which is enough to read a single paint layer in cross section, a single fibre, a single ink line on a document or a single particle in a heterogeneous sample. The IR microscope is what converts FTIR from a bulk technique into a microspectroscopy capable of evidence-grade single-particle identification.

The instrument landscape at Indian SFSLs centres on four vendors. The Bruker ALPHA II is a compact bench ATR-FTIR widely deployed because of its size, speed and price, and the Bruker TENSOR is the larger research-grade unit. Agilent Cary 630 is the other compact ATR favourite. PerkinElmer Spectrum Two and Shimadzu IRTracer-100 round out the routine choices. CFSL Chandigarh runs a Bruker ALPHA II with an IR microscope for paint and fibre work. FSL Sector 14 Madhuban uses ATR-FTIR for daily white-powder triage. CFSL Pune handles explosive-residue ATR-FTIR for the western India workload. NIPER Mohali keeps a pharmaceutical ATR-FTIR with a curated bulk-drug library for counterfeit pharmaceutical investigations.