FT-NIR Analysis of the Hock Process for the Production of Phenol and Acetone

Significance of the topic

Real-time monitoring of the Hock process (cumene oxidation and acid cleavage to produce phenol and acetone) is critical for safety, product quality, and process efficiency. Cumene hydroperoxide (CHP) is both the key oxidation product and the reactive intermediate for cleavage; its concentration must be tightly controlled to avoid off-spec product, hazardous accumulations, and unnecessary downtime. Implementing fast, robust analytical techniques at-line or in-line enables immediate corrective actions, reduces laboratory load, and improves yield and safety in large continuous-flow chemical plants.

Objectives and overview of the study

This application note evaluates Fourier transform near-infrared spectroscopy (FT-NIR) for rapid quantitative monitoring of CHP and cumene in the industrial phenol/acetone (Hock) process. The aims were to demonstrate spectral markers for CHP and cumene, develop chemometric calibrations suitable for in-line use, compare analysis times with titration-based wet chemistry, and show practicality of hazard-certified process FT-NIR analyzers for continuous real-time control.

Methodology

• Sample presentation: Fiber-optic process probe with 1 mm optical pathlength to avoid total absorption at lower-frequency CHP bands and to enable in-line measurement.
• Spectral acquisition: 32 co‑averaged scans at 8 cm^-1 resolution (≈20 s per spectrum); background scan taken with clean, dry probe air to remove probe-environment contributions.
• Spectral processing: Second-derivative and Norris derivative filter (segment length = 11, gap = 0) to enhance small spectral features and remove baseline shifts.
• Chemometrics: Partial least squares (PLS) regression models established for CHP and cumene using selected regions; cross-validation (leave‑out/segment cross-validation) and PRESS analysis to select optimal factor number.
• Calibration ranges: CHP 0–80% (w/w); cumene 6–100% (w/w).

Instrumentation used

• Thermo Scientific Antaris EX (and MX referenced as options) FT‑NIR process analyzer – explosion- and corrosion-rated enclosure for in-line use.
• Fiber-optic probe (1 mm pathlength) for immersion/inline process measurement.
• FT‑NIR acquisition parameters: 32 co‑averaged scans, 8 cm^-1 resolution, 20 s scan time per spectrum.

Key spectral observations

• CHP exhibits distinct OH-related absorption features around 6800 cm^-1 and 4800 cm^-1; these bands are absent in cumene spectra.
• Cumene shows differentiating features in the 6000–5600 cm^-1 region compared with CHP.
• Second-derivative processing accentuates subtle peak-shape differences (notably 6000–5600 cm^-1), improving selectivity in multicomponent calibration.

Main results and discussion

• PLS calibration performance:
  • CHP: Calibration RMSEC = 0.169% (w/w); RMSECV = 0.362% (w/w); model used 3 PLS factors.
  • Cumene: Calibration RMSEC = 0.323% (w/w); model used 4 PLS factors; validation samples matched the calibration accuracy.
• Low RMSECV and small number of PLS factors indicate robust models that capture concentration-related spectral variance without overfitting to irrelevant spectral noise.
• Analysis time comparison: traditional titration (including sample handling) ≈30 minutes; laboratory NIR ≈5 minutes; in-line FT‑NIR ≈20 seconds—demonstrating orders-of-magnitude improvement in responsiveness.
• PRESS plots reached minima with few factors, supporting model simplicity and stability for routine prediction of unknown process samples.

Benefits and practical applications

• Rapid, simultaneous monitoring of multiple process points (oxidation towers, distillation columns) enables a process snapshot for control rooms, supporting trending, alarms, and closed-loop control.
• In-line FT‑NIR reduces off‑spec production, frees analytical staff from routine titrations, and shortens feedback cycles for process optimization.
• Hazard-certified analyzer enclosures facilitate safe deployment in explosive, corrosive, or wet environments typical of petrochemical plants.
• Return on investment arises from reduced laboratory costs, fewer out-of-spec batches, and improved process efficiency.

Future trends and potential applications

• Integration with process control systems for automated feedback and advanced control strategies (model-predictive control) using streaming FT‑NIR data.
• Expansion to multi-analyte, multi-point networks using simultaneous analyzers (e.g., Antaris MX/EX) to provide plant-wide compositional maps in real time.
• Further robustness improvements via adaptive calibrations, transfer learning between probes, and inline standardization to accommodate process drift and probe fouling.
• Application of higher-speed detectors and optimized chemometric workflows to reduce spectrum acquisition time below current 20 s while maintaining accuracy.

Conclusion

FT‑NIR with appropriate probe hardware and chemometric modeling provides accurate, precise, and fast monitoring of CHP and cumene in the Hock process. Distinct spectral markers and carefully chosen derivative preprocessing permit low-error PLS models with few factors, suitable for in-line deployment. Transitioning from titration-based sampling to continuous FT‑NIR yields major time savings and enables real-time process control that enhances safety, reduces waste, and improves product quality.

References

• Heil C. FT-NIR Analysis of the Hock Process for the Production of Phenol and Acetone. Thermo Fisher Scientific Application Note 51711, 2008.