A supplier’s certificate of analysis may show acceptable purity while leaving an important question unanswered: what else is present, and was the method capable of finding it? A well-designed chemical synthesis impurities third party testing program helps technical evaluators verify supplier claims and investigate risks that routine release testing may overlook. Its value depends less on the number of instruments used than on whether sampling, analytical selectivity, and detection limits match the sourcing decision.

For API and fine-chemical procurement teams, the challenge is practical. A trace contaminant can affect patient safety, downstream reaction performance, product stability, or regulatory acceptance. Independent testing should therefore produce more than a pass/fail statement. It should establish what was measured, how reliably it was measured, and which uncertainties remain.

Start with the impurity question, not the instrument list

“Test for impurities” is not a sufficiently defined laboratory request. Chemical synthesis creates several distinct impurity populations, and each requires a different analytical strategy. Before commissioning work, evaluators should review the synthesis route, raw-material specifications, catalysts, solvents, purification steps, and relevant storage conditions.

The resulting risk assessment should distinguish between:

• Process-related organic impurities: unreacted starting materials, intermediates, side-reaction products, and reagent-derived contaminants.
• Degradation products: compounds formed through exposure to heat, moisture, light, oxygen, or incompatible packaging.
• Residual solvents: solvents introduced during reaction, extraction, crystallization, cleaning, or drying.
• Elemental impurities: catalyst residues and metals introduced through raw materials, equipment, or utilities.
• Potentially mutagenic impurities: substances requiring targeted assessment because ordinary related-substances testing may lack sufficient sensitivity or specificity.

Water, inorganic salts, and insoluble residues may also matter, although they often require separate tests. A high HPLC area-purity result does not account for every component of a material. Compounds without a suitable detector response, nonvolatile residues, and water can remain outside that measurement.

The specification must also reflect intended use. An API, an agrochemical intermediate, and a biochemical processing reagent do not automatically share the same acceptance criteria. For industrial materials, even a toxicologically unremarkable impurity may poison a catalyst or interfere with a downstream separation.

Matching analytical methods to the suspected contaminants

Independent laboratories typically combine methods because no single platform covers all impurity classes. The following comparison is a starting point, not a substitute for demonstrating performance in the actual material.

HPLC: useful separation does not guarantee complete coverage

HPLC remains central to related-substances testing because it can separate many nonvolatile organic compounds from a concentrated main component. Technical reviewers should examine resolution at critical peak pairs, integration rules, sample stability, and detector suitability.

Area normalization is particularly easy to overinterpret. A chromatogram reporting 99.8% main-peak area does not necessarily establish 99.8% mass purity. Impurities can have different response factors, and some components may not appear at the selected wavelength. Where quantitative accuracy matters, impurity standards or justified relative response factors should support the calculation.

Photodiode-array peak-purity assessment can help identify certain co-elution problems, but it is not definitive proof that a peak contains only one compound. Orthogonal chromatography or mass spectrometric investigation may be necessary when a potentially consequential impurity could be hidden beneath the principal peak.

GC and LC-MS: trace analysis needs matrix control

Headspace GC often suits residual-solvent testing because it reduces direct introduction of nonvolatile sample components. However, solvent release into the headspace depends on the sample matrix, diluent, incubation temperature, and equilibration conditions. Calibration and recovery studies should represent those conditions.

GC-MS adds identification support, while LC-MS/MS is often selected for targeted trace organic contaminants, including some potentially mutagenic impurities. Neither platform is inherently reliable simply because it is sensitive. Matrix-induced suppression or enhancement, carryover, and extraction losses can dominate the result. Appropriate internal standards, spike recovery, blanks, and dilution studies help reveal these effects.

ICP-MS: distinguish elemental content from chemical form

ICP-MS provides sensitive measurement of many elements, but the result generally describes elemental concentration rather than chemical species. It does not, by itself, identify the oxidation state or organometallic form of a contaminant.

Sample digestion must recover the elements of interest without introducing contamination or losing volatile analytes. Laboratories should demonstrate suitable recovery and control relevant spectral interferences. ICP-OES may be adequate at higher concentrations; selecting ICP-MS solely for its lower instrumental detection limit can add complexity without improving the decision.

Detection limits: the number must belong to the sample

The limit of detection (LOD) describes a level at which an analyte can be distinguished from background using defined criteria. The limit of quantification (LOQ) describes a level at which it can be quantified with demonstrated, acceptable performance. These are not interchangeable, and neither is automatically the same as the laboratory’s reporting limit.

A reporting limit may be set at or above the LOQ for operational or contractual reasons. Consequently, “not detected,” “below LOQ,” and “less than the reporting limit” can communicate different analytical findings. Reports should define these terms rather than expecting procurement teams to infer their meaning.

There is no universal detection limit for “HPLC,” “GC-MS,” or “ICP-MS.” Performance depends on the analyte, detector, preparation procedure, matrix, instrument conditions, and acceptance criteria. A vendor’s instrument sensitivity claim cannot establish the sensitivity of the full analytical method.

A simple dilution calculation changes the interpretation

Suppose a laboratory dissolves 0.50 g of material to a final volume of 50 mL. If the demonstrated LOQ in that prepared solution is 0.20 µg/L, the corresponding sample-basis LOQ is:

(0.20 µg/L × 0.050 L) ÷ 0.50 g = 0.020 µg/g.

This conversion assumes no further dilution or concentration step and adequate recovery. Additional dilution would raise the sample-basis LOQ. Incomplete extraction, matrix interference, or analyte instability could also undermine the claimed sensitivity unless addressed through method performance studies.

Evaluators should therefore request limits expressed in the original material, such as µg/g or percentage by mass, alongside the preparation details. Comparing a solution concentration directly with a solid-material specification is a preventable but consequential mistake.

How low is low enough?

The LOQ should be suitable for the relevant reporting threshold or acceptance limit, with enough margin to support a defensible decision. A method that only quantifies reliably at the specification boundary offers little room for uncertainty or routine variability.

Consider a hypothetical impurity limit of 1 µg/g and a laboratory reporting limit of 5 µg/g. A result reported as “<5 µg/g” cannot establish compliance with the 1 µg/g requirement. Conversely, forcing every method toward the lowest technically achievable limit may increase cost and contamination sensitivity without addressing a meaningful risk.

Near a decision boundary, measurement uncertainty and the agreed decision rule deserve explicit discussion. Where applicable, the laboratory should explain how uncertainty is treated when issuing a conformity statement.

Which standards should guide the testing program?

For pharmaceutical materials, several ICH guidelines provide relevant frameworks. ICH Q3A and Q3B address impurities in new drug substances and products within their respective scopes. Their reporting, identification, and qualification thresholds should not be treated as universal limits for every chemical or pharmaceutical context.

ICH Q3C addresses residual solvents, while ICH Q3D provides a risk-based framework for elemental impurities in drug products. Translating permissible daily exposure values into API or component controls requires consideration of daily intake, route of administration, and contributions from other components—not simply adopting a generic ppm limit.

ICH M7 addresses assessment and control of DNA-reactive mutagenic impurities. The familiar threshold-of-toxicological-concern approach is not a blanket limit for every impurity; compound-specific information, exposure duration, and high-potency categories can alter the assessment. Nitrosamine risks require attention to current, jurisdiction-specific regulatory guidance.

ICH Q2(R2) and Q14 support analytical procedure validation and development. Relevant pharmacopoeial methods and regional requirements may also apply. For nonpharmaceutical fine chemicals, contractual specifications, sector-specific rules, and downstream process requirements often determine the testing framework.

ISO/IEC 17025 accreditation is useful evidence of laboratory competence, but evaluators must inspect its scope. Accreditation for one technique, analyte group, or matrix does not establish coverage for an unrelated test. It also does not, by itself, establish suitability for pharmaceutical GMP release activities.

What to request before accepting a third-party report

Independence matters, but it cannot compensate for an unsuitable sample or weak documentation. A laboratory may analyze an aliquot impeccably while that aliquot fails to represent the shipment.

For chemical synthesis impurities third party testing, the commissioning package should define sampling responsibility, batch identity, container selection, storage conditions, chain of custody, and any retained samples. Composite sampling can obscure localized contamination; individual-container testing may be warranted where heterogeneity is a credible risk.

The technical evidence package should include:

• Method identity and suitability: procedure version, target analytes, matrix, and relevant validation or verification evidence.
• Quantitative basis: calibration approach, reference-standard traceability, response factors, units, and calculation basis.
• Performance at the required level: selectivity, recovery, precision, range, and sample-basis LOQ.
• Run controls: blanks, system suitability, quality-control samples, carryover checks, and relevant acceptance criteria.
• Reviewable results: chromatograms or appropriate raw-data extracts, integration explanations, deviations, and clear treatment of unidentified peaks.

When supplier and independent results disagree, investigate before declaring either laboratory wrong. Differences in moisture correction, preparation, integration, reference standards, or sampling can explain apparently conflicting numbers. Orthogonal testing may resolve a persistent discrepancy more effectively than repeatedly running the same method.

The procurement decision should reflect what the evidence actually proves

A targeted panel confirms only the analytes and conditions within its scope. An untargeted screen may reveal unexpected features, but it does not establish that every possible impurity is absent. Unknown peaks may require identification, quantitative follow-up, and toxicological or process assessment before a sourcing decision is secure.

For technical evaluators across the API, biochemical, and fine-chemical supply chains covered by AgriChem Chronicle, the strongest question is not “Did the material pass?” It is “Was this testing program capable of detecting the failures that matter for our intended use?”

That question links supplier transparency to analytical reality. A defensible program combines representative sampling, risk-based analyte selection, matrix-demonstrated sensitivity, applicable standards, and reviewable evidence. “Not detected” then becomes a properly bounded conclusion—not a promise of absolute purity.