Phenol-formaldehyde (PF) resins are one of the longest-used polymers condensation polymers in industry — binders for plywood and OSB boards, high-pressure laminates (HPL), varnishes, foundry core sands, and adhesives for the woodworking and construction industries. The synthesis reaction proceeds between phenol and formaldehyde in an acidic or alkaline environment; the molar ratio of substrates and the precise termination point of condensation determine both the properties of the final product and the content of unreacted substances: free phenol i free formaldehyde. These two parameters require precise and continuous monitoring throughout the entire reactor operation — for both quality and process safety reasons.
At Gekko Photonics, we design and manufacture process Raman analyzers in Poland — in inline, at-line, and portable variants — and we verify their performance under conditions close to production. In this article, we explain why inline monitoring of free phenol and free formaldehyde is critical for the PF process, and how Raman spectroscopy with an immersion probe eliminates the fundamental limitations of traditional analytical methods.
PF resins — resoles and novolacs, one quality control problem
The two main types of phenol-formaldehyde resins differ in their condensation mechanism and critical quality parameter:
- Resoles (basic catalyst, molar ratio F:P > 1): self-curing without a hardener; key control parameter — free formaldehyde, because its excess generates emissions during thermal processing of end products.
- Novolacs (acidic catalyst, F:P < 1): thermoplastic, cured with hexamine; key parameter — free phenol, because it affects the mechanical properties of the cured material and VOC emissions during hot processing.
Acceptable specification ranges depend on the customer and product standard, but typical industry requirements are:
- Free phenol: 0.3–5.0% (w/w)
- Free formaldehyde: 0.1–3.0% (w/w)
Traditional analytical methods — potentiometric bromination (for phenol), Nash acetylacetone method (for formaldehyde), gas chromatography, or HPLC — require manual sample collection and 30–90 minutes of laboratory analysis. In a batch reactor with a typical cycle time of 3–5 hours, this means a maximum of 3–4 measurement points per batch. This is far too few to react to deviations in real time and control the condensation endpoint.
Safety and regulations — why inline monitoring is a necessity
Phenol is an acutely toxic substance with a high skin absorption rate; the TLV (ACGIH) is 5 ppm, and phenol is listed on the SVHC list under REACH. Any manual sampling from a phenolic resin reactor during synthesis exposes the operator to vapors and the risk of direct skin contact.
Formaldehyde is classified by IARC as a Group 1 carcinogen; the Polish NDS (Maximum Allowable Concentration) is 0.37 mg/m³ (0.3 ppm). The binding occupational exposure limit (BOEL) of 0.3 ppm was introduced in the EU by Directive 2019/983/EU (third amendment to CMRD 2004/37/EC) — effective in all member states since 2021, with the transition period for the healthcare and funeral sectors lifted in July 2024.
Inline monitoring via a Raman immersion probe completely eliminates the need to open the reactor and take samples during synthesis: measurement occurs through a sapphire or quartz window built into the reactor wall, without operator contact with the process medium. Results go directly to the control system — no risk, no delay.
How Raman spectroscopy works in a PF resin reactor
Raman spectroscopy measures inelastic photon scattering on chemical bonds. Each component of the liquid reaction mixture has a characteristic spectral profile:
- Free phenol: strong aromatic ring breathing band at ~1,000 cm⁻¹, C=C stretching bands at ~1,600 cm⁻¹, and an O–H band at ~3,650 cm⁻¹. The ~1,000 cm⁻¹ band is diagnostic and well-resolved even in a dense resin matrix.
- Free formaldehyde: in aqueous solution, it exists mainly as methanediol (HOCH₂OH); diagnostic C–O and C–H bands in the ~900–1,050 cm⁻¹ range.
- Water: characteristic broad O–H band at ~3,400 cm⁻¹; included in the chemometric model as a separate component.
- PF condensate matrix: broad bands of oligomers and polymers form a spectral background, which the PLS model treats as a common component and separates from the signals of active substances.
The immersion probe with optical fiber is mounted in a DN 50 flange or ¾″ NPT port of the reactor — leak-free, with no need to stop the process during installation (for reactors operating at atmospheric pressure). The spectrometer with an external 785 nm or 1,064 nm laser and CCD or InGaAs detector collects a spectrum every 30 seconds to 2 minutes. Near-infrared (NIR) excitation minimizes background fluorescence — particularly problematic in organic resins with 532 nm excitation.
A detailed comparison of hardware configurations for different process types can be found in the article on inline process analyzers.
Chemometrics — calibration and accuracy in practice
The PLS (Partial Least Squares) model is calibrated on a set of samples taken from the actual process or prepared in the laboratory, covering the full working range of concentrations and process conditions (temperature, viscosity, batch composition). Typical calibration parameters:
- Free phenol: calibration range 0.2–6.0% (w/w), RMSECV < 0.15%
- Free formaldehyde: range 0.05–3.5% (w/w), RMSECV < 0.10%
Reference values during calibration: GC or brominometric titration performed in an accredited laboratory. After cross-validation (leave-one-out or k-fold) and residual checking, the model is uploaded to the analyzer software. Results appear in real time on PROFIBUS, PROFINET, or GSM communication interfaces and integrate with any DCS, SCADA, or recipe management system.
When changing raw material suppliers or modifying the formulation: model update takes 1–2 working days — simply add new samples to the calibration matrix and recalculate the regression. No equipment replacement or recertification is required.
More about chemometric methods used in process analyzers is discussed in the article on machine learning in process chemometrics.
Gekko Photonics solutions for PF resin monitoring
At Gekko Photonics, we offer three implementation paths, tailored to the installation architecture and budget requirements:
- Spectrally™ X1 INLINE — analyzer with an immersion probe mounted directly in the reactor or recirculation pipeline. Continuous measurement without sampling, with optional ATEX configuration (available upon request). Optimal solution for batch and continuous reactors when closed-loop control or automatic detection of the condensation endpoint is required.
- Spectrally™ X1 LAB — stationary analyzer for measurements of manually collected process samples (batch control, model validation, calibration work, recipe verification). Used where inline probe installation is impractical — e.g., specific pressure/temperature conditions, limited sample volume, or when the quality team prefers to conduct analysis in a control room next to the production line.
- Spectrally™ X1 PORTABLE — portable analyzer for mobile model verification in the field, rapid identification of phenol and formalin raw material batches upon delivery, and support for the plant laboratory.
Each implementation includes a probe matched to the reactor port, a spectrometer with laser and detector, software with a ready communication interface, a PLS model calibrated on samples from the customer's process, and operator team training. Full product line overview: Gekko Photonics process analyzers.
FAQ — monitoring free phenol and formaldehyde in PF resins
Why is monitoring free phenol so important in PF resin production?
Excessively high free phenol levels reduce the mechanical and emission properties of the final product (plywood, HPL laminates), cause exceedances of VOC emission standards during thermal processing, and may lead to non-compliance with the declaration of performance. Excessively low free phenol in novolac may indicate incomplete condensation. Inline monitoring detects deviations within minutes — before the batch reaches the finished product tank.
Does Raman spectroscopy work in dense, viscous PF resin?
Yes. The probe with a sapphire or quartz window penetrates the liquid medium to a depth of 5–20 mm; dynamic viscosity up to several thousand mPa·s is not a limitation. Excitation at 785 nm or 1,064 nm eliminates most background fluorescence typical of phenolic condensates — a common problem with 532 nm excitation. For very dark resins (high color index), we recommend 1,064 nm.
How long does it take to calibrate a chemometric model for PF resins?
Typically, 30–50 samples representing the full process range are needed: different substrate concentrations, synthesis stage temperatures, and possibly different raw material batches. With samples available, calibration and validation take 3–5 working days. On-site validation (comparison of analyzer results with the reference method under process conditions) — 1–2 days.
What monitoring solutions for PF resins does Gekko Photonics offer?
At Gekko Photonics, we supply Spectrally™ X1 INLINE and X1 LAB analyzers with full implementation for PF resin reactors: immersion probe, PLS model calibrated on samples from the specific process, integration with DCS via industrial standards (PROFIBUS, PROFINET, GSM). Test measurement on customer samples is performed within 10 working days from their delivery. Details at the process analyzers page.
At Gekko Photonics, we select the probe type and laser configuration for the specific PF resin reactor — resole or novolac, pressurized or atmospheric, batch or continuous. We perform an initial test measurement on samples from your process within 10 working days from the moment of their delivery. Contact us via the contact page — we will arrange a 30-minute consultation with an application engineer to specify the installation parameters and the scope of the test measurement.