FTIR Qualification and Performance Verification

FTIR qualification should demonstrate that the configured system can support its intended measurements. A passing transmission check does not, by itself, cover an ATR accessory, sample contact or an identification library. Start by defining the measurement mode, useful spectral region, acquisition settings and decision the result will support; then select references and tests that address those requirements.

This article offers a GuideGxP planning framework for QC laboratories, not a ready-approved qualification protocol. Its matrix and simulated case help separate optical performance, accessory behaviour and method suitability. Acceptance criteria, test frequency and release authority belong in the laboratory’s controlled procedure, with the applicable compendial edition and method requirements identified before execution.

Define the system, accessories and intended use

Describe the system as an instrument, software, detector, optical configuration and identified accessories. Record which combinations are authorised. A laboratory that uses both a transmission film holder and attenuated total reflection (ATR) has more than one measurement configuration, even when both share the same spectrometer and workstation. Microscopy or diffuse reflection adds further interfaces and sampling questions.

Specify the result: comparison with a reference spectrum, identity decision, investigation of an unexpected material or quantitative determination. These activities need different evidence. For example, an exploratory comparison may help an investigation but cannot automatically release a raw material. A qualified spectral range also needs to cover the bands that the particular procedure relies on.

Build a requirements-to-evidence list before accepting a service report. Identify the configuration tested, functions covered, exclusions, reference status and records retained. If the report concerns only the base optical bench, record how the installed accessory and laboratory procedure will be assessed separately. A service certificate is useful evidence, but its title does not establish the scope of its conclusions.

EDQM Annex 4 distinguishes periodic or motivated checks from in-use checks within the OMCL framework [1]. Applying that guidance outside its network requires a justified assessment; it is not a universal pharmaceutical tolerance table.

Control acquisition conditions and the background

An FTIR spectrum depends on how both the background and sample were acquired. Define the background configuration, accessory state, acquisition sequence and conditions that require renewal. Retain enough information to determine whether an unexplained feature already existed in the background or appeared with the sample. Simply collecting a new background after an anomaly can hide a useful comparison if the original record is discarded.

Record spectral range, nominal resolution, number of scans, apodisation, relevant detector settings and any atmospheric compensation or baseline processing. Apodisation is a mathematical weighting used when transforming the acquired signal; it is part of the measurement context, not a cosmetic display choice. A denser displayed data grid does not, on its own, demonstrate better optical resolution.

Consider temperature, humidity, purge condition and the time needed for the configured instrument to stabilise. Water vapour or carbon dioxide features may suggest a mismatch between acquisitions, but their appearance alone does not locate the cause. Compare the sequence, background and environmental records before attributing the observation to a sample impurity or instrument failure.

For a meaningful repeatability experiment, decide what is being repeated. Reprocessing one file evaluates processing consistency. Reacquiring without moving the specimen includes acquisition variation. Removing and replacing an ATR specimen adds contact and positioning variation. Use the comparison that answers the qualification question and label the evidence accordingly, without imposing one universal replicate count.

Choose performance checks that answer a defined question

Wavenumber-scale verification addresses the location of spectral features; resolution addresses the ability to distinguish nearby features; noise characterisation addresses variation that may obscure useful information. Passing one check cannot establish the others. Define the evaluation method, units, acquisition conditions and decision rule for each test before reviewing the result.

For noise or signal-to-noise comparisons, record the spectral region, signal definition, noise statistic and acquisition settings. Peak-to-peak and root-mean-square measures are different quantities. Comparing values produced with different settings or definitions can create an apparent deterioration or improvement unrelated to the question being investigated. A software report should expose enough detail to make that distinction.

The following original matrix links a configuration to a specific evidence claim. It contains planning examples rather than prescribed limits. Select additional checks when the method, instrument design or risk assessment requires them; omit a proposed check only with an explanation of why its failure mode is outside the authorised use.

Keep failed or ambiguous observations in the report. A pass/fail summary should be traceable to the spectrum, settings, calculation and approved criterion. If the test setup was wrong, document the reason and the corrected experiment; do not silently replace the unsuccessful run.

Select references for their actual scope and condition

Choose a reference for the property and measurement mode it supports. A material useful for locating bands is not automatically certified for absorbance, resolution, sample-contact performance or library discrimination. Reference identity, lot or unit, certificate version, storage, condition and authorised use should be available at the point of testing.

The NIST SRM 1921b transmission certificate distinguishes band evaluation methods and acquisition conditions. The live certificate read for this article was issued on 15 September 2026, with validity through 31 December 2027 [2]. Check its current status before use.

Do not transfer a transmission reference value to an ATR measurement merely because the polymer name is identical. Establish whether the reference and evaluation are appropriate for that optical mode. Likewise, a damaged film or contaminated reference cannot be rescued by using its historic certificate as if its physical state were unchanged.

Separate uncertainty from tolerance. The uncertainty attached to a certified value describes the reference information; the acceptance criterion expresses the performance needed for the intended use. Their relationship belongs in the decision rule. If the available reference cannot support the required decision with suitable confidence, improve the reference or test design rather than widening the criterion after seeing the result.

Before testing, resolve missing information: applicable edition, relevant bands, evaluation algorithm, certificate conditions and authorised handling. An attractive spectrum and a familiar reference name do not replace these records. Keep a controlled link between the reference used and the resulting qualification dataset.

Assess the accessory and sampling contribution

ATR introduces a contact interface. Crystal condition, specimen surface, placement and the pressure mechanism can affect the acquired spectrum. Transmission has its own preparation issues, including film thickness, cell path and preparation consistency. Qualification planning should distinguish a stable optical system from a repeatable way of presenting the actual material to it.

Use this accessory checklist before interpreting a changed spectrum:

• Confirm accessory identity, crystal material and the authorised configuration.
• Check cleanliness, visible damage and compatibility with the sample and cleaning procedure.
• Record specimen condition, contact surface, placement and permitted pressure setting.
• Confirm the background and acquisition settings match the intended method.
• Retain the original and processed spectra, including any ATR correction.
• Assess repositioning separately from repeated acquisition at one position.

Simulated case — spectrum changes after ATR replacement. A QC laboratory replaces an ATR accessory. The same retained polymer specimen then gives different relative band intensities and a poorer library match. A transmission reference check passes. This supports the tested transmission performance; it does not establish that the replacement ATR produces equivalent sampling conditions.

The investigation first confirms accessory identity and settings, then checks the clean background and a suitable accessory reference. A planned repositioning comparison shows variable intensity when contact is inconsistent. Standardising the permitted presentation improves repeatability, but a systematic difference from the old library remains. The team therefore evaluates accessory and processing compatibility with the identification method before authorising routine use.

This case contains no measured data or universal pressure recommendation. It illustrates two separate questions: whether the new configuration performs as intended, and whether the existing identity procedure remains suitable. Do not increase force beyond authorised limits, polish a crystal or alter internal optics to obtain a better match. Such actions can damage the accessory or create a safety risk.

Separate qualification results from identification-library decisions

Instrument qualification supports confidence in the measurement system. A library comparison asks whether a sample can be identified under a particular procedure. An instrument can pass its performance checks while the library lacks the correct material, physical form or acquisition mode. Conversely, a familiar sample can match a library despite an unresolved qualification problem.

Record library version, reference provenance, acquisition mode, preprocessing, comparison algorithm and authorised acceptance logic. Establish how near matches, unexpected bands, mixtures and materials outside the library are handled. The highest score among available entries is a ranking result, not proof that the correct identity is represented or that alternative identities have been excluded.

Jung and colleagues demonstrated an ATR identification workflow for environmental polymers [5]. Its relevance here is the need to evaluate identification in a defined application; its material set and decision criteria are not pharmaceutical release criteria.

For QC adoption, challenge the intended procedure with representative materials and plausible confusions. Include variation that routine sampling is expected to encounter. Decide when an inconclusive spectrum requires investigation or another suitable technique. Avoid changing smoothing, spectral regions or match thresholds after examining a difficult sample without an approved, documented rationale.

Keep the records connected: a released instrument configuration, an authorised method, a controlled library and the analyst’s review of the actual sample. None is a substitute for the others. A qualification report should explicitly state what it does not establish about the identity method.

Maintain the qualified state through routine use and change

Translate the approved qualification into an operating plan: checks before use, periodic review, reference management, maintenance and actions after a failure. Choose frequency from the application, stability history, workload and consequences of an undetected problem. A calendar alone does not address changes between scheduled checks.

Accessory replacement, detector or source work, relocation, software updates and changes to processing may trigger different reassessments. Start from the affected function and the evidence that may no longer apply. Do not assume that every change demands the entire original qualification, or that a maintenance visit automatically restores every authorised application.

For an anomaly, preserve spectra, backgrounds, settings, reference records and the sequence of actions. Define any temporary restrictions on use and evaluate potentially affected results from the last defensible evidence of acceptable performance. Where the issue concerns a reportable pharmaceutical result, follow the laboratory’s investigation and quality decision process rather than repeatedly measuring until a favourable spectrum appears.

Return to service should identify the resolved cause or justified disposition, successful relevant checks, remaining limitations and approval responsibility. Review trends under comparable conditions, including changes in background or sampling repeatability. The useful endpoint is a clear statement of which FTIR configurations and procedures may be used, with evidence that remains retrievable when the next change occurs.

Sources and applicability

The matrix, checklist and case are GuideGxP editorial recommendations. EDQM [1] is OMCL guidance. USP ⟨854⟩ was consulted through its public introduction only; the full applicable USP and Ph. Eur. 2.2.24 texts must be checked by the laboratory. No inaccessible compendial limits are reproduced. The image is illustrative.

1. EDQM / OMCL. PA/PH/OMCL (18) 24 R1. Qualification of Equipment, Annex 4: Qualification of IR Spectrophotometers (effective September 2019).
2. NIST. SRM 1921b: Infrared Transmission Wavelength/Wavenumber Standard. Certificate issued 15 September 2026.
3. USP. ⟨854⟩ Mid-Infrared Spectroscopy (2019), public introduction.
4. NIST. Fourier Transform Infrared Spectroscopy. Tools & Instruments (updated 2019).
5. Jung MR et al. Validation of ATR FT-IR to identify polymers of plastic marine debris, including those ingested by marine organisms. Marine Pollution Bulletin 127 (2018), 704–716. doi:10.1016/j.marpolbul.2017.12.061.

Related Pharma Lab articles

• Analytical instrument qualification and USP ⟨1058⟩
• Qualification, method validation and system suitability
• UV-Vis qualification: wavelength and photometric performance