Process control of liquid-phase phenol chlorination is dominated by the need to suppress local thermal feedback loops that shift the product distribution from 2-chlorophenol toward 2,4-dichlorophenol and higher chlorinated homologues. In a production-scale bubble column or stirred batch reactor the chlorination sequence proceeds through parallel and consecutive electrophilic substitutions; each chlorine addition liberates hydrogen chloride and transfers substantial heat to the reaction mass. The temperature field is non-uniform at the chlorine sparger and at the baffle-to-wall transition, where low fluid velocity permits hot spots to persist. A temperature measurement located only in the bulk phase can lag these hot spots by 5 °C to 15 °C depending on circulation and bubble-driven turbulence, so reactor temperature control must be treated as a distributed thermal regulation problem rather than a single-point set-point exercise. The selectivity relationship between 2-chlorophenol and 2,4-dichlorophenol is therefore dominated by the ability to remove heat at the point of chlorine dissolution while maintaining a stable bulk temperature within the narrow band required for mono-chlorination. Jacketed glass-lined reactors, external circulation loops, and feed pre-cooling are employed in combination because no single heat-exchange surface can be relied upon when the chlorine feed rate is modulated to maintain bulk temperature.

The Selectivity Between 2-Chlorophenol and 2,4-Dichlorophenol Depends on More Than the Chlorine Feed Ratio

Temperature exerts a direct effect on the rate constants of the competing reaction paths, and chlorine feed ratio alone cannot compensate for a poorly controlled thermal profile. In the first substitution, phenol is attacked at ortho and para positions; the ortho adduct leads to 2-chlorophenol, while the para adduct leads to 4-chlorophenol. The second substitution on 2-chlorophenol is typically accelerated relative to the first substitution because the ring is activated by the hydroxyl group and the first chlorine modifies the electron density, but the temperature sensitivity of the second substitution differs from that of the first. Because each step follows an Arrhenius-type rate equation with different apparent activation energies, a temperature increase raises the ratio of 2,4-dichlorophenol formation to 2-chlorophenol formation when the activation energy for the second chlorination exceeds that of the first under the selected solvent and catalyst conditions. Published kinetic data for these specific isomers are scattered across solvent systems; therefore, the exact temperature coefficient must be obtained by reaction calorimetry in the actual process fluid rather than transferred from aqueous chlorination data. The practical consequence is that a set-point change of a few degrees can shift the product split when the process is operated near the maximum allowable 2-chlorophenol concentration. In a batch reactor, the instantaneous concentration ratio is further complicated by the fact that the local concentration of dissolved chlorine in the boundary layer around the gas bubble is not equal to the bulk concentration. This means that even a perfectly controlled bulk temperature cannot eliminate all consecutive chlorination if the gas-liquid mass transfer is poor and the local chlorine concentration at the gas interface remains high.

During scale-up from a 2 L bench-scale chlorination flask to a pilot-scale 50 L jacketed reactor, the limiting heat transfer area per unit volume decreases sharply. A bench-scale vessel can be immersed in an external bath and maintains a near-isothermal reaction mass, but a pilot-scale vessel must reject heat through a jacket and, if installed, an internal coil. The gas-liquid mass transfer rate at the chlorine sparger changes with impeller type, baffle configuration, and gas hold-up; a Rushton turbine operating at a tip speed below the full turbulent regime can create oxygenated zones near the shaft that are not representative of the bulk temperature. When chlorine gas is introduced through a dip tube, the local heat release at the bubble interface is intense enough to raise the local temperature even when the bulk temperature remains on set point. This local superheat drives the consecutive chlorination to 2,4-dichlorophenol and 2,6-dichlorophenol before the molecules are diluted into the bulk. The reaction mass is therefore not uniformly selective; the bulk temperature reading may correspond to a mixture that is already enriched in dichloro compounds at the feed point. To suppress this distribution, the chlorine feed is often diluted with nitrogen or recirculated off-gas, and the sparger is positioned below the impeller discharge to maximize rapid gas dispersion. A sparger located behind a baffle or near the wall produces a dead zone where the 2-chlorophenol selectivity is lost and where fouling from tarry byproducts accelerates. Published data for this specific gas-liquid configuration is limited because much of the industrial optimization is held as proprietary process knowledge.

What Limits the Cooling Rate When a Batch Temperature Excursion Threatens 2-Chlorophenol Selectivity?

Cooling rate is limited by the heat transfer coefficient of the process-side film, the thermal resistance of the glass lining, the available coolant flow, and the time constant of the temperature sensor. In a glass-lined steel jacket the overall heat transfer coefficient is often below 400 W m⁻² K⁻¹ under viscous or fouling conditions; the glass layer contributes a conductive resistance that cannot be eliminated by increasing coolant flow alone. When a batch temperature excursion begins, the jacket outlet temperature may remain low while the bulk temperature rises because the heat flux has not yet penetrated the vessel wall; a thermocouple placed in the jacket loop can therefore mislead the cascade controller if it is used as the primary feedback. The maximum safe cooling rate is further constrained by the temperature of the heat transfer fluid and the potential for thermal shock to the glass lining. A sudden demand for cold water at 5 °C after a period of steam heating can generate tensile stress in the glass and cause fracture, especially if the vessel has been in service for several years. Process control is therefore designed to limit the rate of utility change rather than maximize it. In chlorination systems that can accumulate unreacted chlorine in the headspace, a cooling failure scenario is evaluated using adiabatic calorimetry and pressure relief sizing according to API 520 and API 521; the relief rate is determined by the reaction runaway rather than by external fire alone. The temperature control loop must therefore work within the thermal inertia of the vessel and the mechanical limits of the glass, not simply against the chemistry.

Within a continuous reactor train consisting of a jacketed CSTR followed by a plug-flow finishing stage, the temperature control strategy shifts from batch heat removal to feed-forward disturbance rejection. The phenol feed is preheated or pre-cooled to a defined inlet temperature, while chlorine is injected through a mass flow controller with an upstream pressure regulator selected to prevent condensation in the rotameter. The CSTR temperature is controlled by a cascade loop in which the primary controller uses the reactor temperature and the secondary controller uses the jacket inlet temperature or the thermal fluid flow. A split-range valve arrangement supplies low-pressure steam to the jacket during initial heat-up and chilled water during the chlorination exotherm. The transition between heating and cooling is a critical operating boundary: when the control valve crosses the split range, a dead band or valve stiction can cause the reactor temperature to oscillate by more than ±1 °C. Because the 2-chlorophenol-to-2,4-dichlorophenol selectivity is sensitive to temperature history, even short-lived oscillations can broaden the product distribution. In this configuration, the jacket supply temperature is measured by a platinum resistance thermometer with an accuracy of ±0.15 °C in the operating range, and the reactor temperature is measured by a thermocouple in a thermowell designed for vortex-induced vibration using ASME PTC 19.3 TW-2010.

When a Jacketed CSTR Is Operated Below the Threshold at Which 2,4-Dichlorophenol Formation Accelerates

Operating below the temperature threshold does not remove the need for tight control because the reaction rate is still exothermic and the threshold is a kinetic boundary, not a physical phase transition. The apparent activation energy for the formation of 2,4-dichlorophenol may be higher than that for 2-chlorophenol, so a reduction in set point can selectively suppress the second chlorination. However, lowering the set point also reduces the solubility of chlorine in the organic phase and can lower the reaction rate to a point where the control system operates with excessive integral action. When the reaction rate is low, the temperature controller may keep the jacket fully closed for an extended period; the integral term winds up, and when the exotherm eventually begins, the jacket cannot open quickly enough to prevent an overshoot. This overshoot drives the mixture across the threshold, and the same integral action then persists in the cooling direction, causing a symmetrical excursion below set point. The result is a temperature cycle that repeatedly crosses the selectivity boundary. This behavior is controlled by external reset feedback or by disabling integral action during the idle period, not by reducing the proportional band alone. A temperature-controlled process mass flow controller for chlorine can also be used to reduce the magnitude of the disturbance; the chlorine feed is ramped according to the heat removal capacity, and the temperature controller is allowed to trim only a small portion of the reaction enthalpy.

Jacket Duty, Circulation Rate, and Thermowell Placement

Jacket duty is determined by the product of the heat transfer area, the overall heat transfer coefficient, and the differential temperature between the reaction mass and the coolant. In a baffled reactor the process-side heat transfer coefficient depends on the impeller Reynolds number, the Prandtl number of the reaction mass, and the fouling resistance associated with chlorinated phenolic tar. Increasing the agitation speed improves heat transfer but also increases the gas hold-up and may produce a vortex that draws chlorine into the headspace. The circulation rate through the jacket is selected to maintain a coolant outlet temperature no more than 5 °C below the coolant inlet temperature under maximum heat load; higher coolant flows reduce the differential temperature and lower the thermal shock risk but increase the pressure drop through the jacket circuit. Thermowell placement determines whether the measured temperature represents the bulk, the impeller discharge, or a stagnant zone. A thermowell installed downstream of the baffle may report a temperature that lags the true reaction temperature by several minutes, so the control loop must be derated to avoid instability. Thermowells in gas-liquid chlorination service are subject to vortex-induced vibration and corrosion; the wake frequency calculation and Strouhal number evaluation follow ASME PTC 19.3 TW-2010, and the thermocouple element is selected from IEC 60584-1 tolerance Class 1. In installations where the process fluid is aggressive, a tantalum or PTFE sheathed thermowell is used, but the slower thermal response of the non-metallic sheath must be compensated by a lower proportional gain.

In an online analyzer loop, a slipstream is quenched with cold solvent to arrest the reaction and then injected into a gas chromatograph configured for phenol, 2-chlorophenol, 2,4-dichlorophenol, and 2,6-dichlorophenol quantification. The analyzer cycle time of 15 min is too slow for closed-loop temperature control, but it is used to correct the selectivity model and set the compensating trim on the temperature controller. The temperature controller remains the primary protection against consecutive chlorination because it can act within a few seconds, whereas analyzer feedback is limited by the sample transport delay and the column retention time. In a production line where the set point is fixed by the analytical target, the operator adjusts the chlorine feed ratio only after the measured 2-chlorophenol concentration falls outside the specification window. The temperature controller then compensates for changes in coolant temperature, exothermic reaction rate, and heat of dilution. Historical alarm reviews from three continuous campaigns indicate that the majority of selectivity deviations are preceded by a period of poor jacket temperature control rather than by a sudden change in feed ratio.

Thermowell Selection in Chlorination Service Is Governed by Wake Frequency and Material Compatibility

Thermowell design cannot be separated from temperature control quality because the sensor insertion point determines the dead time and the susceptibility to vibration-induced noise. In a chlorination reactor, the process-side fluid is often a two-phase mixture of organic liquid, aqueous acid, and chlorine gas; the thermowell encounters fluctuating lift and drag forces as gas bubbles pass. The wake frequency calculation from ASME PTC 19.3 TW-2010 is not a formality: a thermowell installed at the impeller discharge can operate near its natural frequency and transmit a high-amplitude oscillation to the temperature element, which the controller then interprets as a real temperature disturbance. The resulting valve movement is unnecessary and can introduce additional thermal cycles. Material selection for the thermowell is also constrained by the corrosive environment. Stainless steel may be acceptable in the gas phase but is often attacked by wet hydrogen chloride in the liquid phase, so PTFE or tantalum sheaths are used. A non-metallic sheath increases the thermal response time constant of the temperature loop, and the derivative action must be removed or filtered to avoid amplifying the lag. The thermowell root diameter, bore diameter, and insertion length are selected so that the first natural frequency exceeds the vortex shedding frequency by a margin specified in the design standard; in practice this may limit the maximum permissible tip velocity to a value below the normal agitation speed.

When the Heat Transfer Fluid Supply Temperature Is Allowed to Cycle Across the Split-Range Boundary

Split-range control is used to operate the jacket with steam during heat-up and cooling water during the exotherm, but the transition zone is a known source of temperature disturbances. A valve positioner with a poorly characterized mechanical linkage may stick at 20 % to 30 % of travel, preventing the jacket from switching smoothly from steam to cooling water. During this period the reactor temperature can rise without effective counteraction because the steam valve has not fully closed and the cooling water valve has not fully opened. The temperature controller responds by increasing the control signal, but the valve does not move until the stiction is overcome. When the valve finally moves, it opens too far, and the reactor temperature undershoots. This behavior is worsened by the interaction between the split-range controller and the cascade jacket temperature loop because the jacket loop can interpret the sudden opening of the cooling water valve as a new set-point error. Corrective measures include using a single high-performance globe valve with a positioner compliant with ISA-75.01.01, applying a small dead-band compensation in the positioner, and specifying a minimum transition time between heating and cooling modes. The thermal shock concern is not eliminated by control tuning; the glass-lined vessel still requires a controlled ramp rate in the jacket supply temperature, typically limited by the vessel manufacturer to a specified difference between the metal jacket and the glass surface.

In a chlorination unit where the ratio of 2-chlorophenol to 2,4-dichlorophenol is monitored by offline sampling, the temperature controller set point may be moved only after the analytical result is confirmed. The laboratory method is typically a gas chromatographic procedure with a flame ionization detector, calibrated against reference standards for each chlorophenol isomer. The sample is quenched immediately after withdrawal with a bicarbonate solution to neutralize dissolved hydrogen chloride and prevent continued chlorination in the sample container. The reported precision of the analytical method is often ±0.2 % absolute for major components, but the sampling error in a two-phase reactor can exceed the instrument error. The temperature controller cannot correct for poor sampling because the sample point may be located in a region that is not representative of the bulk. To address this, a sample port is placed downstream of the impeller discharge, and the sample line is maintained at the reactor temperature to avoid condensation. The loop is arranged so that the temperature controller receives a filter-averaged signal from three temperature elements located at the bottom, mid-point, and top of the liquid phase; the median value is selected to reject a single failed sensor. This sensor voting arrangement is implemented in the safety system and in the basic process control system, with the safety interlock set at a temperature above the normal operating band but below the decomposition or runaway threshold.