Nuclear Magnetic Resonance (NMR) Spectroscopy

A nucleus has spin only if its proton-plus-neutron count gives a non-zero spin quantum number. Hydrogen-1 has spin 1/2. Carbon-13 has spin 1/2. Carbon-12, the dominant isotope, has spin zero and is completely NMR invisible. This is the first reason carbon NMR is harder than proton NMR: only 1.1 percent of natural carbon is the magnetically active isotope.

In a strong external field B0, spin-1/2 nuclei align either with the field (lower energy) or against it. The resonance frequency at which an RF photon flips the spin is given by nu = gamma·B0/2π. For 1H in an 11.7 T magnet that works out to 500 MHz, which is why a "500 MHz NMR" is shorthand for a magnet with that field strength.

The RF pulse hits the sample for a few microseconds. After the pulse, the perturbed magnetisation precesses in the transverse plane and induces a tiny voltage in the receiver coil. That voltage decays as the spin system relaxes (T1 longitudinal, T2 transverse), giving a free induction decay. A Fourier transform converts the time-domain decay into the conventional frequency-domain spectrum.

Sensitivity scales with the cube of the gyromagnetic ratio and with natural abundance. 1H is the easiest nucleus to observe because gamma is highest and abundance is essentially 100 percent. 13C is roughly 5,700 times less sensitive than 1H, which is why a 1H spectrum needs 16 scans on a milligram sample (under five minutes) and 13C on the same sample often needs 1,024 scans run overnight.

Every nucleus in a molecule sits in a slightly different electronic environment. Electrons around a nucleus shield it from the external field, so a nucleus surrounded by electron density experiences a slightly weaker effective field and resonates at a slightly lower frequency. A nucleus near an electronegative atom or a pi system has its electrons pulled away (deshielded) and resonates at a higher frequency. The chemical shift axis maps these tiny frequency differences in parts per million relative to a reference, conventionally tetramethylsilane (TMS) at 0 ppm.

The following ranges carry most of the diagnostic information in routine 1H interpretation.

The 13C axis covers 0 to 220 ppm because 13C shifts are more sensitive to electronic environment than 1H. Aliphatic carbons sit at 10 to 50 ppm, carbons bonded to oxygen at 60 to 90, aromatic and alkene carbons at 110 to 150, and carbonyl carbons at the deshielded extreme (esters and amides 165 to 175, ketones and aldehydes 195 to 210).

The standard interpretive approach is to read chemical shift first and use it to assign a functional class. A peak at 9.5 ppm in 1H is almost always an aldehyde; a peak at 200 ppm in 13C is almost always a ketone carbonyl. Shift narrows the possibilities to a handful before multiplicity and integration finish the job.

If two magnetic nuclei sit on adjacent carbons, they sense each other through bonds. Each neighbour's spin (with or against the field) shifts its partner's resonance by a few Hertz, splitting one peak into a multiplet whose pattern encodes how many neighbours the proton has.

For first-order systems the n+1 rule applies. A proton with n equivalent neighbours splits into n+1 lines: no neighbour gives a singlet, one a doublet (1:1), two a triplet (1:2:1), three a quartet (1:3:3:1), with intensities following Pascal's triangle. The classical case is the ethyl group (CH3-CH2-X) where the CH3 appears as a triplet and the CH2 as a quartet. The triplet-plus-quartet pattern of the ethyl group is among the most reliably identified signatures in routine 1H NMR interpretation.

The spacing between adjacent lines is the coupling constant J in Hz. The same J appears on both sides of a coupled pair, which is how you confirm two multiplets are actually coupled rather than coincidentally close. J also encodes geometry: vicinal H-H couplings (3J) follow a Karplus dihedral-angle dependence, axial-axial couplings in cyclohexanes are large (10 to 12 Hz) and axial-equatorial small (2 to 4 Hz), aromatic ortho couplings are 7 to 9 Hz, meta 1 to 3 Hz and para 0 to 1 Hz. A trained eye reads ring substitution patterns from J alone.

Integration is the third leg of 1H NMR interpretation. The area under each peak (or each multiplet, treated as one peak) is proportional to the number of equivalent protons giving rise to it. Integration is reported as a stepped curve overlaid on the spectrum or as a numeric value per peak. The ratios are what matter: a peak integrating for 3 next to a peak integrating for 2 is a CH3 and a CH2 in the same molecule. Absolute integration values are arbitrary and are normalised against the smallest assigned peak.

The combined reading of chemical shift, multiplicity, integration and coupling constant is what extracts a structure from a 1H spectrum. Chemical shift narrows the functional environment. Multiplicity counts the neighbours. Integration confirms how many protons in each environment. Coupling constants confirm geometry and connectivity. Four readings, one spectrum, one structure.

A modern NMR spectrometer is built around a superconducting magnet kept at liquid helium temperature inside a vacuum-insulated cryostat, with a liquid nitrogen jacket to slow helium boil-off. A 400 MHz instrument runs at 9.4 T, a 600 MHz at 14.1 T, and the 1000 MHz instruments at top international labs at 23.5 T. Price tracks field strength: a 400 MHz routine instrument is around fifty lakh rupees, a 600 MHz research instrument runs into a few crores.

The RF probe sits in the bore at the field centre. The sample, dissolved in a deuterated solvent in a 5 mm tube (or 3 mm for limited sample), drops in via a spinner that rotates the tube at about 20 Hz to average out radial field inhomogeneity. The probe holds the RF coils that send the pulse and receive the FID.

Three subsystems keep the spectrum stable. The lock channel continuously monitors the 2H resonance of the deuterated solvent and feeds back to the field regulator, correcting any drift in real time. This is why every NMR sample needs a deuterated solvent: it supplies the lock signal and stays invisible in the 1H spectrum. The shim coils are small electromagnets around the probe that homogenise the field across the sample volume to a few parts per billion; a well-shimmed magnet gives baseline-resolved peaks under 1 Hz wide. The console handles RF generation, pulse sequencing and ADC, with workstation software (TopSpin, Delta, MestReNova) controlling the experiment and processing the FID.

Sample requirements for routine 1H are 5 to 50 mg in 0.5 to 0.6 mL of solvent, filtered to remove particulates, with no paramagnetic impurities (trace Fe, Cu, Mn destroy resolution). 13C and 2D experiments need the upper end of that range.

Routine 1D experiments include 1H, 13C with proton decoupling, 19F, 31P and DEPT 135 or DEPT 90 to distinguish CH3, CH2, CH from quaternary carbons. The 2D set adds COSY (1H-1H scalar coupling), HSQC (1H to directly bonded 13C), HMBC (1H to 13C two and three bonds away), NOESY (through-space) and TOCSY (total correlation within a spin system).

Indian forensic NMR has grown sharply since about 2018, driven by the surge in novel psychoactive substances and the recurring problem of adulterated traditional-medicine products.

NPS characterisation is the flagship application. Synthetic cannabinoids (JWH, AM, AB-FUBINACA, ADB-PINACA series), synthetic cathinones (mephedrone, MDPV, alpha-PVP), fentanyl analogues and tryptamines arrive as small white powders that GC-MS triages by molecular ion. The problem is that minor structural isomers give nearly identical mass spectra. NMR resolves this by reading the substitution pattern directly: a 1H spectrum of AB-FUBINACA shows the indazole NH, fluorobenzyl CH2, valine alpha-H and t-butyl methyls at distinct shifts and integrations; a substituted analogue swaps one signal for another at a predictable ppm. CFSL Hyderabad commissioned a Bruker AVANCE NEO 400 MHz in 2022 specifically for NPS confirmation, and the throughput is now a steady stream of seized-powder cases from the southern states.

Seized-cannabis profiling needs both quantitation and structural confirmation. 1H NMR distinguishes THC from CBD and CBN by their olefinic and aromatic proton patterns, and quantitative NMR (qNMR) with an internal standard such as maleic acid measures cannabinoid content without a calibrated reference for each cannabinoid. Terpene fingerprints in the upfield aliphatic region encode geographic origin and chemovar. CSIR-IICT Hyderabad and a few state SFSLs with SAIF access use this for ganja and charas profiling.

Ayurvedic adulterant analysis is a significant casework category. AYUSH-licensed formulations are repeatedly found to contain allopathic NSAIDs (diclofenac, ibuprofen, paracetamol), corticosteroids (dexamethasone, prednisolone), anti-diabetics (glibenclamide, metformin) or sildenafil analogues. 1H NMR of an extract gives the adulterant away: a paracetamol acetyl singlet at 2.05 ppm and the para-substituted AA'BB' aromatic pattern at 7.0 and 7.4 ppm leaves no doubt. NIPER Mohali and CDRI Lucknow handle most of this referral work.

Counterfeit pharmaceutical structural confirmation runs alongside. When a state drug controller seizes a suspect batch under the Drugs and Cosmetics Act, NMR confirms whether the API matches the label and whether any structurally related impurity, process intermediate or substitution is present. NIPER Mohali's central instrumentation handles pharmaceutical structure elucidation including HSQC and HMBC experiments.

Drug-impurity profiling can sometimes link two seizures to the same clandestine source by matching the minor-impurity 1H fingerprint, although GC-MS impurity profiling is usually more sensitive. Polymer characterisation is shared with FTIR: 1H NMR of a polymer in CDCl3 or DMSO-d6 quantifies copolymer ratios, identifies plasticisers and reads polypropylene tacticity from methyl multiplets, none of which FTIR does quantitatively. Edible-oil authenticity is an FSSAI-driven application where 1H NMR of mustard, olive, coconut, groundnut and palm oils flags adulteration of premium oils with cheaper substitutes at the few-percent level, and the same approach catches honey adulterated with high-fructose corn syrup.

Reading a 1H spectrum in practice follows a fixed sequence. Count the distinct peak groups to get the number of unique proton environments. Read each integration to get the relative number of protons in each environment. Read each chemical shift to assign a functional environment. Read each multiplicity to count adjacent protons. Read each coupling constant to confirm geometry. Combine the four readings and a structure usually falls out within fifteen minutes for a small organic.

DEPT 135 and DEPT 90 are the two standard editing experiments for 13C. DEPT 135 makes CH3 and CH peaks point up, CH2 peaks point down, and quaternary carbons disappear entirely. DEPT 90 leaves only CH peaks visible. A combined reading distinguishes the four carbon classes (CH3, CH2, CH, Cq) in one experiment and is the single biggest interpretive shortcut for 13C work, because quaternary carbons (no attached protons) are otherwise hard to identify and assign.

The limitations of NMR are real. Sensitivity is poor: routine 1H needs about 1 mg, full 2D characterisation 5 to 50 mg, against the nanogram quantities a triple-quadrupole MS handles. For trace residue work (postmortem tissue, swab, ng-level seizure) MS is the only realistic option. Cost is high: a 400 MHz runs around fifty lakh rupees and a 600 MHz three to five crores, with annual cryogen costs of five to ten lakh and a magnet quench another five to ten lakh to recover from. Acquisition is slow: a 1H scan takes 5 to 30 minutes including lock and shim, and a full 2D set (COSY, HSQC, HMBC, NOESY) takes overnight. Quaternary carbons give weak 13C signals and DEPT cannot detect them, so HMBC is the standard workaround at the cost of more acquisition time.

Benchtop NMR is changing some of this. Permanent-magnet instruments at 60, 80 or 90 MHz from Magritek, Nanalysis and Anasazi cost roughly thirty lakh, fit on a bench, need no cryogen and run on a wall socket. Resolution is poorer (overlap is heavier, second-order coupling more common, 13C barely usable), but for routine identification, qNMR of pure compounds and teaching they are very capable. A handful of Indian SFSLs and university programmes have started procuring them alongside a central superconducting magnet.

The Indian NMR map clusters around a few vendors and central facilities. Bruker (AVANCE NEO 400, AVANCE III HD 600, AVANCE NEO 800) dominates research, with Jeol (JNM-ECZ400R, ECZ500R) the second choice. Major access points are the IISc Bangalore NMR Centre (700 and 800 MHz), IIT Bombay and IIT Madras SAIFs (400 and 500 MHz open access), NIPER Mohali (400 and 600 MHz for pharma), CDFD Hyderabad (600 MHz for structural biology), CSIR-IICT Hyderabad, CDRI Lucknow (400 and 600 MHz for natural products and drug development) and the CFSL Hyderabad Bruker 400 commissioned in 2022 for NPS forensic casework. Several state SFSLs route NMR work to the nearest SAIF rather than maintaining their own magnet.