NMR structure elucidation of impurities means using one- and two-dimensional nuclear magnetic resonance experiments (typically 1H, 13C, DEPT, COSY, HSQC and HMBC), supported by high-resolution mass spectrometry, to assign the full chemical structure of an unknown impurity or degradation product. It is the step that turns "unknown peak at RRT 1.24" into an identified impurity with a defined structure, which is what ICH Q3A(R2) and Q3B(R2) expect once an impurity exceeds its identification threshold. This guide explains when identification is required, how to obtain a sample that is fit for NMR, which experiments answer which questions, a worked example, and where quantitative NMR (qNMR) fits in.
When an unknown impurity must be identified
For new drug substances, ICH Q3A(R2) links the obligation to identify an impurity to the maximum daily dose. Above the identification threshold, the applicant is expected to characterise the structure of the actual impurity; if identification proves impossible, a summary of the unsuccessful laboratory work should be included in the application.
| Maximum daily dose | Reporting threshold | Identification threshold | Qualification threshold |
|---|---|---|---|
| ≤ 2 g/day | 0.05% | 0.10% or 1.0 mg/day intake (whichever is lower) | 0.15% or 1.0 mg/day intake (whichever is lower) |
| > 2 g/day | 0.03% | 0.05% | 0.05% |
Source: ICH Q3A(R2), Attachment 1. Lower thresholds can be appropriate for unusually toxic impurities.
A quick worked example shows why the "whichever is lower" clause matters. At a maximum daily dose of 500 mg, 0.10% corresponds to 0.5 mg/day, so the 0.10% threshold applies. At 1.5 g/day, 0.10% would be 1.5 mg/day, which exceeds 1.0 mg; the effective identification threshold therefore drops to 1.0 / 1500 = approximately 0.067%. For drug products, ICH Q3B(R2) sets dose-dependent identification thresholds for degradation products that range from 1.0% (or 5 µg total daily intake) for doses below 1 mg down to 0.10% for doses above 2 g.
Beyond the Q3 thresholds, a confident structure is also the starting point for a mutagenicity assessment under ICH M7(R2): you cannot run a structural-alert evaluation on a peak you have only characterised by retention time and nominal mass.
Getting a sample that is fit for NMR
LC-MS can often propose a structure directly from a mixture, but NMR usually needs the impurity isolated, or at least strongly enriched. Most of the time spent on an elucidation project goes here, not on the spectrometer.
Enrichment strategies
- Mother liquors and process streams: process impurities are frequently concentrated in crystallisation filtrates.
- Forced degradation: a degradant seen at 0.12% on stability may form at several per cent under targeted stress (see our guide to forced degradation studies).
- Preparative or semi-preparative HPLC: transfer the analytical method to a volatile mobile phase (formic acid, ammonium formate or ammonium acetate rather than phosphate) so fractions can be dried without leaving buffer salts.
- Targeted synthesis: when a structure can be proposed from MS data, synthesising the candidate is often faster than isolating it, and gives material for a reference standard.
Before the sample goes into the magnet
- Re-inject the isolated fraction on the original HPLC method to confirm retention time, purity and that the compound has not changed during work-up (TFA, heat and concentration can all cause artefacts).
- Confirm the accurate mass and isotope pattern by LC-MS to fix the molecular formula before interpreting any NMR data.
- Choose the deuterated solvent deliberately. DMSO-d6 dissolves most APIs and slows exchange of NH and OH protons so they can be observed and correlated; CDCl3 can be slightly acidic and may degrade acid-labile compounds; CD3OD and D2O exchange away labile protons.
- Remove residual isolation solvents thoroughly; their signals overlap analyte regions and complicate integration.
Which NMR experiments answer which questions
No single experiment solves a structure. The skill lies in choosing the smallest set that removes ambiguity for the specific question, which is usually "how does this impurity differ from the parent molecule?"
| Experiment | What it tells you | Typical use in impurity work |
|---|---|---|
| 1H | Number of proton environments, integrals, multiplicity, coupling constants | First comparison with the parent API; spotting lost or new protons |
| 13C with DEPT or APT | Carbon count and CH3/CH2/CH/quaternary editing | Confirming the carbon skeleton; carbonyls and quaternary carbons |
| 19F, 31P | Heteronuclear environments with little background | Fluorinated or phosphorus-containing APIs; quick impurity fingerprinting |
| COSY / TOCSY | Proton-proton coupling networks (spin systems) | Tracing alkyl chains, rings and sugar units |
| HSQC | One-bond 1H-13C correlations | Assigning protonated carbons without a full 13C spectrum |
| HMBC | Two- and three-bond 1H-13C correlations | Joining fragments across quaternary carbons and heteroatoms; locating substituents |
| NOESY | Through-space proximity of protons | Regio- and stereochemistry, E/Z isomers, epimers |
When material is scarce, proton-detected 2D experiments (HSQC and HMBC) are generally more sensitive than a directly detected 13C spectrum, so carbon chemical shifts can often be read from the 2D data instead.
Fix the molecular formula by accurate-mass MS first, then use NMR to answer one question: what changed relative to the parent? Compare every spectrum of the impurity side by side with the API acquired under the same conditions, and let HMBC and NOESY settle positional and stereochemical questions that mass spectrometry cannot.
A practical elucidation workflow with worked example
- Establish the molecular formula from accurate mass and isotope pattern; calculate the double-bond equivalents (rings plus double bonds).
- Acquire 1H of both impurity and API in the same solvent; tabulate shifts and integrals side by side.
- Count carbons and their multiplicities from 13C/DEPT or from HSQC.
- Build spin systems with COSY or TOCSY.
- Connect the spin systems through quaternary carbons and heteroatoms using HMBC.
- Use NOESY for relative configuration or regiochemistry where needed.
- Check that the proposed structure accounts for every signal, every correlation and the molecular formula, and that it is chemically plausible for the process or degradation pathway.
- Where possible, confirm by synthesis and matching of retention time and spectra; this also yields material for a characterised standard.
Worked example: a +16 Da oxidative degradant
Consider an API containing an N-methylpiperidine ring and a para-disubstituted benzene ring. Peroxide stress produces a degradant whose protonated molecular ion is 16 Da heavier, consistent with addition of one oxygen. MS alone leaves at least two candidates: an N-oxide or a hydroxylated aromatic ring.
- N-oxide: the aromatic region keeps its four-proton AA'BB' pattern, while the N-CH3 singlet and the ring CH2 groups next to nitrogen move noticeably downfield (the N-methyl typically shifts by around 1 ppm), with matching changes in the attached carbons in HSQC.
- Aromatic hydroxylation: the aromatic integral falls from four to three protons and the pattern changes to that of a 1,2,4-trisubstituted ring (a small meta coupling and a larger ortho coupling become visible), while the N-methyl signal is essentially unchanged. HMBC from the remaining aromatic protons to the new oxygen-bearing carbon fixes the position of the hydroxyl group.
The two hypotheses are distinguished by a 1H spectrum and an HSQC, without the need for extensive data. This is typical: a well-chosen comparison with the parent often solves the structure faster than a full de novo assignment.
Quantitative NMR (qNMR) for impurity standards
Once an impurity is identified, it usually needs a reference standard with an assigned content. Chromatographic purity alone ignores water, residual solvents, inorganic salts and non-chromophoric impurities. qNMR addresses this because, under properly chosen acquisition conditions, signal area is directly proportional to the number of nuclei, regardless of the molecule. ICH Q2(R2) includes quantitative 1H-NMR with an internal standard as one of its illustrative examples of analytical procedures.
With an internal standard of known purity, the purity of the analyte is calculated as:
Px = (Ix / Istd) × (Nstd / Nx) × (Mx / Mstd) × (mstd / mx) × Pstd
where I is the integral, N the number of protons giving rise to the signal, M the molar mass, m the weighed mass and P the purity. Practical points that decide whether the result can be trusted:
- Select a certified, traceable internal standard whose signal does not overlap any analyte or impurity signal and which is stable in the chosen solvent.
- Use a relaxation delay long enough for full recovery of magnetisation; a common rule is at least five times the longest T1 of the signals being integrated.
- Choose analyte signals that are representative, well resolved and free from exchange; avoid NH, OH and signals near the water or solvent peaks.
- Weigh on a suitably calibrated microbalance; weighing is frequently the largest contributor to uncertainty.
- Apply careful phasing and baseline correction, and integrate consistently across replicate preparations.
More on what a standard's certificate should contain is in our impurity reference standards buyer's guide.
Common pitfalls and a reporting checklist
Pitfalls
- Isomers with identical mass: regioisomers and epimers are invisible to MS; only NMR (often NOESY or HMBC) or comparison with an authentic sample resolves them.
- Rotamers and tautomers: doubled or broadened signals may indicate restricted rotation rather than a mixture; variable-temperature spectra or a different solvent can clarify.
- Artefacts from isolation: TFA adducts, salt-form changes or degradation during drying can produce a structure that is not the impurity actually present in the batch.
- Exchangeable protons: in CD3OD or D2O, NH and OH signals disappear, which can be misread as a missing functional group.
- Residual solvent and water peaks misassigned as analyte signals.
What a structure elucidation report should contain
- Source of the sample, isolation route and chromatographic purity of the isolate.
- Accurate-mass data, proposed molecular formula and mass error.
- All NMR spectra with acquisition parameters, solvent and reference.
- A full assignment table (1H, 13C, key COSY, HMBC and NOESY correlations) for both impurity and parent.
- The reasoning that excludes alternative structures, and a proposed formation pathway.
- Confirmation by synthesis or co-injection, where performed.
How SDS Labs can help
SDS Labs operates a 400 MHz FT-NMR spectrometer offering 1H, 13C, 31P, 19F, DEPT, APT and qNMR, together with COSY, HSQC, HMBC, NOESY and TOCSY, in a range of deuterated solvents. Our NMR service is used for structure elucidation and confirmation of APIs, intermediates and impurity reference standards, alongside LC-MS and mass spectrometry, and our custom synthesis team can prepare proposed impurity structures for confirmation.
To discuss an unknown impurity, share your chromatograms and any MS data through our contact page, call +91 90295 20144 or email info@sdslabs.in.
Frequently asked questions
Why is NMR needed if LC-MS already gives the molecular mass?
Mass spectrometry gives the molecular formula and fragments, but many impurities share the same mass as other plausible structures. Regioisomers, epimers and N-oxide versus hydroxylated products can have identical masses. NMR shows which atoms are connected and where substituents sit, through experiments such as HSQC, HMBC and NOESY, so it is usually needed to confirm a structure with confidence.
How much material is needed for NMR structure elucidation?
It depends on the molecule, the field strength and the experiments required. A proton spectrum can be obtained on small amounts, while proton-detected 2D experiments such as HSQC and HMBC need more, and a directly detected carbon-13 spectrum needs the most. Where material is scarce, laboratories rely on HSQC and HMBC to read carbon shifts instead of a full carbon spectrum.
At what level must an unknown impurity be identified under ICH?
Under ICH Q3A(R2), for new drug substances with a maximum daily dose up to 2 g, impurities above 0.10% or 1.0 mg per day intake, whichever is lower, should be identified. Above 2 g per day the identification threshold is 0.05%. ICH Q3B(R2) sets separate dose-dependent identification thresholds for degradation products in drug products.
Which 2D NMR experiments are most useful for impurity identification?
HSQC assigns each proton to its directly attached carbon, COSY and TOCSY trace coupled proton networks, and HMBC links fragments across quaternary carbons and heteroatoms, which is often decisive for locating a new substituent. NOESY shows which protons are close in space and is used to resolve regiochemistry and stereochemistry, such as distinguishing epimers or E/Z isomers.
What is qNMR and why is it used for impurity standards?
Quantitative NMR measures content by comparing the integral of an analyte signal with that of a certified internal standard of known purity. Because signal area is proportional to the number of nuclei, it does not require a reference standard of the impurity itself. Its result inherently accounts for water, solvents and salts that chromatographic purity misses, making it well suited to assigning content to newly identified impurity standards.
References
- ICH Q3A(R2): Impurities in New Drug Substances
- ICH Q3B(R2): Impurities in New Drug Products
- ICH Q2(R2): Validation of Analytical Procedures (EMA, Step 5)
- ICH M7: Assessment and Control of DNA Reactive (Mutagenic) Impurities (EMA)
- USP General Chapter <761> Nuclear Magnetic Resonance Spectroscopy (proposed revision)