Automated quality control of gasoline using NIR spectroscopy

Importance of the topic

Gasoline quality control requires reliable, rapid and cost-efficient methods to determine octane performance (RON, MON, AKI), key compositional parameters (oxygenates, benzene, olefins, aromatics) and physical properties (density). Conventional reference methods (CFR engine tests, gas chromatography, density meters) are accurate but time-consuming, resource-intensive and require skilled operators. Implementing near-infrared spectroscopy (NIRS) with robust chemometric models enables simultaneous, rapid screening of multiple parameters, reduces laboratory workload, and supports high-throughput quality assurance and process control applications.

Objectives and overview of the study

The application note investigates the feasibility of using the OMNIS NIR Analyzer Liquid to predict multiple gasoline quality attributes in a single measurement. Primary goals were to develop and validate chemometric models for RON, MON, AKI, density, oxygenates, benzene, olefins and total aromatics, and to demonstrate automated, safe sample handling compatible with ASTM procedures cited in the note (ASTM D8340, D8321, D6122). A calibration/validation strategy (75%/25% split) was used and models were evaluated by standard figures of merit.

Methodology

• Measurement mode: Transmission NIR spectroscopy across 1000–2250 nm at a controlled sample temperature (30 °C).
• Sample handling: Fully automated sampling with OMNIS Sample Robot M – NIR using closed disposable glass vials (8 mm pathlength) to minimize flammability risk.
• Chemometrics: Models built in OMNIS Software; 75% of samples used for calibration and 25% reserved for external validation.
• Reference methods: CFR engine tests for octane metrics (RON, MON, AKI); gas chromatography for olefins, aromatics, benzene and oxygenates; laboratory density meter for density.
• Performance evaluation: Correlation diagrams and figures of merit including R2 (prediction), SEC, SECV and SEP were used to assess model quality and expected routine precision.

Used instrumentation

• OMNIS NIR Analyzer Liquid — near-infrared spectrometer for liquids (1000–2250 nm), temperature-controlled sampling, measurements in under 10 seconds.
• OMNIS Sample Robot M – NIR — automated sample handling for up to 3 racks (up to 231 vials), supports disposable vials (8 mm pathlength) and minimizes operator exposure to flammable fuels.
• Disposable borosilicate vials, 8 mm transmission pathlength and associated vial holder and sample racks.
• OMNIS Software with Quant Development license for calibration model creation and OMNIS stand-alone licensing for instrument control.

Main results and discussion

The developed NIR models show strong correlations with laboratory reference values across the targeted parameters, indicating suitability for routine screening and quality control. Representative figures of merit from the external validation are summarized below:

• AKI: R2P = 0.974; SEP ≈ 0.45 (units consistent with index) — excellent agreement with CFR-derived values.
• MON: R2P = 0.958; SEP ≈ 0.37 — high predictive power for motor octane number.
• RON: R2P = 0.970; SEP ≈ 0.64 — strong performance for research octane number.
• Olefins (mass %): R2P = 0.916; SEP ≈ 1.31% — good prediction of olefin content (GC reference).
• Density (g·cm−3): R2P = 0.991; SEP ≈ 0.0009 g·cm−3 — very high accuracy compared with density meter measurements.
• Aromatics (mass %): R2P = 0.973; SEP ≈ 0.97% — reliable quantification of total aromatics (GC reference).
• Benzene (mass %): R2P = 0.979; SEP ≈ 0.02% — high sensitivity and precision for benzene determination.
• Oxygenates (mass %): R2P = 0.992; SEP ≈ 0.05% — excellent prediction performance for oxygenate content.

Discussion points:

• Speed: Individual NIR measurements are completed in seconds, while delivering simultaneous multi-parameter outputs, providing a major time advantage over sequential wet-chemistry and chromatographic assays.
• Automation and safety: Closed disposable vials and robotic handling reduce manual intervention and flammability risk, enabling high-throughput operation and improved laboratory safety.
• Calibration dependence: As with all spectroscopic models, performance depends on the representativeness of the calibration set; inclusion of diverse fuel matrices and periodic model maintenance are required to preserve predictive accuracy.
• Regulatory alignment: Temperature-controlled measurement and validated models support compliance with referenced ASTM practices and allow rapid screening prior to confirmatory testing when needed.

Benefits and practical applications

• High-throughput quality control in fuel production, distribution terminals and refineries for rapid pass/fail screening or trend monitoring.
• Incoming fuel inspection at blending facilities to verify octane and compositional targets before processing or shipment.
• Process analytical technology (PAT) integration for near-line or at-line monitoring to support timely process adjustments.
• Reduction in consumables, operator time and turnaround compared to multiple reference methods, lowering per-sample analysis cost.

Future trends and opportunities

• Expanded calibration libraries and transfer learning to extend applicability across regional fuel formulations and biofuel blends (e.g., higher ethanol content, oxygenated additives).
• On-line and in-line NIR process analyzers for continuous monitoring and tighter process control in refineries and blending loops.
• Hybrid workflows combining rapid NIR screening with targeted GC or engine testing for regulatory compliance and trace-level analyte confirmation.
• Improved chemometric strategies (robust preprocessing, adaptive calibration, uncertainty quantification) to increase model resilience to matrix shifts and instrument-to-instrument variability.
• Integration with laboratory information management systems (LIMS) and automation platforms to enable automated reporting, trending and decision support.

Conclusion

The study demonstrates that the OMNIS NIR Analyzer Liquid, combined with automated sample handling, can deliver accurate, rapid, multi-parameter gasoline analysis suitable for routine quality control. Models for octane indices, compositional metrics and density achieved high correlation with reference methods and acceptable prediction errors for many industrial QC applications. NIRS is not a replacement for all reference methods where legal confirmation is required, but it is an effective primary screening tool that significantly reduces response time and laboratory burden while improving throughput and safety.

References

• Metrohm Application Note AN-NIR-022: Automated quality control of gasoline using NIR spectroscopy — OMNIS NIR Analyzer Liquid (source document summarized).
• Instrument and accessory configuration details and product numbers referenced in the application note (OMNIS NIR Analyzer Liquid; OMNIS Sample Robot M – NIR; 8 mm disposable vials; OMNIS Software Quant Development license).