Validation Lab
Field spectroscopy is only valuable when backed by defensible reference data. The lab is where that certainty is built.
Why Validation Matters
A spectrum is a pattern. Without rigorous validation against real chemistry, it is just a curve. IDIR’s validation laboratory is the foundation of every model we deploy.
Spectra Must Be Anchored to Chemistry
Calibration models require reference laboratory measurements to assign meaning to spectral features. Without them, predictions cannot be verified or trusted.
Results Must Be Defensible
In regulated industries and commercial disputes, analytical results must be traceable to recognised reference methods and accredited data. ISO 17025 accreditation provides that traceability.
Models Must Be Maintained
Spectroscopic models drift with changes in raw materials, instruments and processes. Ongoing validation against reference data keeps deployments accurate and fit for purpose.
Our Validation Workflow
From parameter definition to documented deployment — a structured six-step process that ensures every model is built on solid scientific foundations.
-
1
Define the Parameter
Agree the target analyte, acceptable accuracy threshold, measurement context (at-line, in-line, field) and any regulatory requirements before any laboratory work begins.
-
2
Collect Representative Samples
Design a sampling plan that captures expected natural variation — across raw material sources, processing batches, seasonal changes and concentration ranges relevant to the application.
-
3
Run Reference Method Analysis
Each sample is analysed using the established reference method (ASTM, ISO or validated in-house). Reference values form the calibration target and provide the traceability chain.
-
4
Capture Spectra
Spectra are recorded under controlled conditions on the target instrument. Replicates, measurement conditions and sample presentation are documented for reproducibility.
-
5
Build and Validate the Model
Preprocessing optimisation, calibration model development, cross-validation and independent test set validation — with full reporting of RMSEC, RMSECV, RMSEP and R².
-
6
Deploy with Documentation
The validated model is deployed to the client instrument or WebGUI platform, accompanied by a full documentation package: calibration report, validation summary and deployment recommendation.
Documentation Outputs
Every validation engagement produces a structured documentation package.
Feasibility Report
Assessment of whether spectroscopy can solve the stated problem — based on initial spectra, literature review and reference data. Go/no-go recommendation with supporting rationale.
Calibration Report
Full calibration dataset, preprocessing steps, model parameters and cross-validation performance statistics — the complete record of how the model was built.
Validation Summary
Independent test set results, prediction accuracy, outlier analysis and fitness-for-purpose evaluation against the agreed accuracy threshold.
Method Transfer Report
Documentation supporting transfer of the validated method to a client's instrument or platform — including instrument equivalence testing and re-validation data if required.
Deployment Recommendation
Practical guidance on measurement frequency, sample handling, instrument maintenance and model monitoring to ensure ongoing performance in the field.
Discuss Validation Requirements
Tell us your application, reference method and accuracy target — we’ll outline a validation approach and timeline.