In the liquid-phase direct chlorination of ethylene to 1,2-dichloroethane (EDC), the circulating reaction medium is not an inert solvent but an active participant in catalyst transport, heat removal, and byproduct accumulation. The specification that water in the recycle EDC be maintained at or below 1 ppm wt is established from production-scale observation that exceeding this concentration for more than a few hours initiates a fault sequence involving ferric chloride hydrolysis, chlorine consumption in side reactions, hydrogen chloride evolution, fouling deposition, and chloride-induced localised corrosion. A typical direct chlorination loop operates a gas-liquid bubble column reactor at 50–70°C and 2.5–4.0 barg with a molar ethylene-to-chlorine ratio of 1.05–1.15:1. The catalyst is ferric chloride at a concentration generally reported in the range of 50–200 mg/kg as FeCl₃ in the liquid EDC phase. Water enters the loop through raw materials, recycled EDC from purification, nitrogen blanketing, and utility interfaces. Discrete liquid sample moisture is measured by coulometric Karl Fischer titration according to ASTM E1064-16 or volumetric titration according to ASTM E203-16; the analytical lower limit of detection for modern coulometric cells is below 1 µg water, allowing reliable quantification at 0.1 ppm levels. The 1 ppm threshold is therefore not an analytical detection limit but an operational control limit. Process data from continuous units show that at water concentrations below 0.3–0.5 ppm, the direct chlorination reaction remains stable with high selectivity to EDC, and the formation of heavy oxygenated compounds is negligible. As the water concentration approaches 1 ppm, the rate of catalyst hydrolysis becomes detectable through an increase in pressure drop across the reactor overhead condenser, a shift in vent gas hydrogen chloride concentration, and a gradual loss of chlorine conversion that cannot be recovered by increasing catalyst addition alone. The fault line is not a single event but a threshold where the net rate of hydrolytic damage exceeds the normal operational capacity for catalyst makeup, dryer regeneration, and corrosion control.

What Is the Operational Significance of 1 ppm Water in Ferric Chloride-Catalysed Direct Chlorination?

The catalytic cycle in direct ethylene chlorination depends on the Lewis acidity of anhydrous ferric chloride, which polarises molecular chlorine and facilitates electrophilic addition to ethylene. Water competes with chlorine for the ferric centre, and even trace hydration alters the speciation of the catalyst in EDC solution. At 1 ppm wt water, the molar ratio of water to ferric chloride in a recycle stream containing 100 mg/kg FeCl₃ is approximately 0.09:1, which is sufficient to hydrolyse a small but kinetically meaningful fraction of the iron complex because the hydrolysis equilibrium constant strongly favours the formation of Fe(OH)Cl₂ and FeOCl species relative to the anhydrous chloride. The immediate consequence is a reduction in apparent turnover frequency of the catalyst. Laboratory kinetic studies in dried EDC indicate that the apparent rate constant for chlorine consumption is first order in dissolved chlorine and first order in ferric chloride at low water content, but the observed rate constant decreases as water concentration rises. Published data for production-scale direct chlorination at 1 ppm water are limited; however, the operational response of industrial reactors is consistent with a partial deactivation mechanism because a given conversion can be restored only partially by injecting additional anhydrous FeCl₃. The injected catalyst itself rapidly picks up water and hydrolyses, so the more effective response is removal of water upstream of the reactor or regeneration of recycle dryers. In addition to catalyst deactivation, water reacts directly with dissolved chlorine to form hydrogen chloride and hypochlorous acid according to Cl₂ + H₂O ⇌ HOCl + HCl. Hypochlorous acid is a chlorohydrinating agent that adds to ethylene to form 2-chloroethanol, which can further react in the presence of HCl and iron species to form chloroacetaldehyde and heavier condensation products. These oxygenated byproducts contaminate the EDC product and can create fouling precursors in downstream distillation columns. The 1 ppm water boundary is therefore operationally significant because it marks the concentration where the combined rates of catalyst hydrolysis, hypochlorous acid formation, and oxygenated byproduct generation become measurable against the normal process variability of ±0.2–0.5% in reactor conversion and selectivity.

From a materials degradation standpoint, the 1 ppm wt water control limit protects carbon steel and austenitic stainless steel surfaces from acidic chloride environments that arise when hydrogen chloride and water co-condense or concentrate at phase boundaries. Dry EDC with less than 0.5 ppm water is comparatively non-corrosive to carbon steel, but the introduction of 1 ppm water together with chlorine-derived HCl creates a condition where liquid films or droplets in vent coolers and reboiler tube sheets can develop acidic chloride concentrations far higher than the bulk dry EDC water concentration. Carbon steel in wet EDC service is then susceptible to general corrosion and hydrogen blistering; austenitic stainless steels such as 316L have improved resistance to uniform corrosion in dry EDC but remain vulnerable to pitting and crevice corrosion in the presence of oxidising ferric ions and chlorides. Standard laboratory corrosion tests according to ASTM G31-72(2021) for immersion corrosion and ASTM G48-11 for pitting and crevice corrosion resistance are used to qualify metallic materials, but the accelerated nature of the test does not reproduce the low water activity of a nominally dry EDC recycle loop. In production-scale forced-circulation reboilers, tube leaks at rolled joints are a common ingress path for water; the resulting local water concentration at the tube face can exceed 10 ppm even when bulk recycle water remains near 1 ppm. The resulting deposit of FeOCl and Fe(OH)₃ on the tube surfaces converts a minor heat transfer derate into an electrochemical corrosion cell, because the deposit traps chloride ions and generates a local acidic environment. Inspection methods qualified under ASTM E384-17 for microindentation hardness are sometimes used to detect hydrogen-induced hardness changes in carbon steel after extended acid exposure, while scanning electron microscopy with energy-dispersive X-ray spectroscopy is used to confirm iron oxychloride and iron oxide deposit composition. The material selection for new equipment in direct chlorination service therefore splits into two zones: anhydrous EDC below 1 ppm water permits carbon steel with corrosion allowance, while potential wet zones such as overhead vapour lines, reboiler condensates, and scrubbing systems require PTFE-lined carbon steel, glass-lined steel, or nickel-based alloys. The operational boundary at 1 ppm water is not solely a chemistry limit; it is also the point at which the material compatibility assumptions in the mechanical design of the loop become invalid.

Thermal Degradation of FeCl₃ Hydrates at Reboiler and Vent Condenser Surfaces

Thermal degradation of ferric chloride hydrates at reboiler and vent condenser surfaces provides a secondary moisture concentration mechanism that can amplify the effect of 1 ppm water in the recycle loop. In a purification column reboiler operating at 150–180°C, dissolved iron chloride species and trace water are exposed to surface temperatures well above the boiling point of EDC at operating pressure. Under these conditions, hydrated ferric chloride species such as FeCl₃·6H₂O and Fe(OH)Cl₂ undergo thermal hydrolysis and release HCl gas according to FeCl₃·6H₂O → Fe(OH)Cl₂ + 2HCl + 4H₂O, with subsequent conversion to FeOCl and Fe₂O₃. The HCl released at the reboiler surface dissolves in the EDC liquid film and returns to the loop, while the iron oxide deposit adheres to the tube wall and reduces the heat transfer coefficient. The standard design fouling resistance for reboilers in EDC purification service is typically set at 0.00017–0.00034 m²·K/W for clean organic fluids; once iron oxide deposits form, the actual fouling resistance can exceed 0.0005 m²·K/W and continue to rise. Vent condenser surfaces face a different degradation mode: low temperatures of 10–30°C allow condensation of water and HCl that are present in the overhead vapour, even when the bulk liquid water concentration is 1 ppm. The resulting acidic condensate can contain HCl at concentrations above 100 ppm by weight, causing rapid aqueous corrosion of carbon steel. Materials for vent condenser construction in direct chlorination service are therefore specified with a corrosion allowance or upgraded to 316L, PTFE-lined carbon steel, or graphite. The combination of high-temperature hydrolysis at the reboiler and low-temperature acid condensation at the vent condenser means that 1 ppm moisture in the recycle liquid does not stay uniformly distributed but is concentrated by thermal operations into more aggressive local conditions.

Because the direct chlorination loop is a closed recycle system with multiple phase changes, water ingress from utility interfaces is rarely visible as a single spike but appears as a gradual upward drift in the online moisture reading that is punctuated by larger deviations during reboiler or traced-line excursions. Common ingress sources on a production-scale unit include steam tracing leaks on EDC transfer lines, tube-to-tubesheet leaks in shell-and-tube reboilers, water-contaminated seal flush from circulation pumps, saturated drying media in the nitrogen blanket header, and incorrect regeneration of molecular sieve dryers. Each source has a distinct signature in process data: a reboiler leak produces a continuous exponential-like rise in recycle water concentration with no relation to feed gas moisture; a leaking steam tracing line produces a diurnal or weather-dependent pattern because tracing loads vary with ambient temperature; a saturated nitrogen dryer produces a plateau that responds to regeneration and then falls. The online moisture analyser in the liquid EDC recycle line is therefore placed downstream of the reactor cooler and before the feed gas injection point to capture water entering from all upstream sources. The analyser sample conditioning system uses a Hastelloy C-276 sample probe, a heated filter to remove iron particulates, and a vaporising membrane to transfer water from the organic phase to a dry nitrogen carrier gas for tunable diode laser absorption spectroscopy detection. The total sample transport delay is kept below 60 s to support closed-loop response. When the recycle water concentration reaches 1 ppm, the control system initiates a source isolation procedure: the steam tracing pressure is reduced, reboiler condensate conductivity is checked, nitrogen dryer bed outlet dew point is verified below −70°C, and a liquid sample is drawn for laboratory Karl Fischer confirmation. Operating experience shows that repeated cycling around 1.0–2.0 ppm without identifying the ingress source leads to a cumulative fouling condition in which the recycle pump discharge pressure rises and the reactor differential temperature across the cooler narrows, even after the water concentration returns to below 0.5 ppm.

When Steam Tracing Leaks Intersect the EDC Recycle Loop

When steam tracing leaks develop on EDC recycle sample lines or on overhead vapour lines maintained above the acid dew point, the resulting water ingress is frequently small and intermittent. Because the tracing line is at 3.0 barg saturated steam and the process line operates at 2.5–4.0 barg, the pressure differential can be low or even negative, so a leak at a flange gasket may not produce a visible steam plume but can still allow water to accumulate in the insulation and diffuse through a pinhole. The water enters the process stream as a liquid film on the inner wall rather than as a uniformly mixed vapour, producing localised FeCl₃ hydrolysis at the wall boundary. A production-scale direct chlorination unit with heat-traced overhead vapour lines can therefore show bulk moisture readings below 1 ppm while the vent condenser inlet nozzle experiences water concentration excursions above 5 ppm, because the water travels as a liquid film and does not fully dissolve before reaching the cooler. The resulting acidic corrosion at the condenser inlet is often detected first by ultrasonic wall-thickness measurements under ASME B31.3 process piping inspection intervals or by a rise in iron concentration in the condensed hydrocarbon layer. The corrective action is to remove the steam tracing from service, inspect the flange gasket, replace with a spiral-wound gasket meeting ASME B16.20, and install a low-pressure dry nitrogen purge on the outer insulation annulus to prevent condensation. This scenario illustrates that the 1 ppm fault line cannot be managed solely by bulk moisture measurement because local ingress mechanisms can create surface-level water concentrations far above the bulk value.

Process control for moisture in direct ethylene chlorination requires an alarm and interlock philosophy that distinguishes between feed gas moisture, liquid recycle moisture, and vent gas dew point. Feed gas moisture is typically measured at the outlet of the ethylene dryer and the chlorine supply vaporiser. Ethylene dryer beds using molecular sieve 3A are designed for an outlet moisture content below 0.1 ppmv under normal conditions, with regeneration initiated at 0.2 ppmv or on a timed cycle of 8–12 h. Chlorine supplies are specified with water content below 50 ppm wt in the liquid phase in some industrial standards, but for direct chlorination, chlorine is vaporised and passed through a mist eliminator before feed; liquid chlorine moisture above 10 ppm is considered a common source of loop water accumulation. The liquid EDC recycle moisture analyser is configured with a pre-alarm at 0.8 ppm, a high alarm at 1.0 ppm, and a high-high alarm at 2.0 ppm. The high-high alarm activates an automatic reduction in chlorine feed rate and initiates a dryer switching sequence. Vent gas dew point is monitored by a chilled-mirror analyser; the acid dew point of the vent gas rises as HCl and water concentrations increase, and an increase in measured dew point from −20°C to −5°C at the given operating pressure indicates moisture or HCl excursion before the liquid recycle analyser responds. The control narrative follows IEC 61511 for safety instrumented functions where the chlorine feed shutdown is assigned a SIL 2 requirement due to the toxic gas release potential of the direct chlorination reactor. Moisture interlocks are not necessarily self-actuating safety functions but are implemented in the basic process control system with independent hardwired high-pressure and high-temperature shutdowns. The effectiveness of the moisture control logic depends on the reliability of the online analyser sample conditioning system; therefore the sample transport line is sloped towards a collection pot and purged with dry nitrogen to prevent liquid water holdup in dead legs. Field experience shows that a sample line dead leg containing only 10 mL of water can produce a delayed moisture response of 2–4 h and cause a false indication of compliance while the actual recycle water concentration has already exceeded 1 ppm.

Feed Gas Drying Specification and Recycle Loop Moisture Analytics

Feed gas drying specification and recycle loop moisture analytics form the measurement and control foundation for maintaining the 1 ppm operational fault line. Ethylene feed gas from a cracker or storage sphere is dried in fixed-bed molecular sieve dryers using 3A or 4A zeolite, with design outlet moisture specified below 0.1 ppmv and regeneration gas heated to 220–260°C. The dryer bed is monitored by pressure drop across the bed and by an online dew point transmitter placed immediately downstream. The ethylene dryer outlet sample line is made of 316L stainless steel with electro-polished internal surface to minimise water adsorption and memory effects. Chlorine feed moisture is controlled at the point of vaporisation by monitoring the liquid chlorine supply specification and by maintaining a slight positive pressure nitrogen purge on the vaporiser vent. Recycle EDC moisture is measured by a slipstream analyser that draws liquid from the circulation pump discharge through a high-pressure sample cooler to reduce the sample temperature to 20–25°C before the Karl Fischer or tunable diode laser absorption interface. The sample is filtered through a 0.5 µm sintered Hastelloy filter to remove iron particles, and the filter is inspected weekly for reddish-brown deposits that indicate iron hydrolysis. Laboratory confirmation of online moisture readings is performed by coulometric Karl Fischer titration according to ASTM E1064-16; the method covers water in organic solvents in the range of 1 mg/kg to 10 g/kg, with a precision of ±5% relative at 1 mg/kg depending on sample handling. Sample containers are glass vials with PTFE-lined caps, purged with dry nitrogen and dried at 105°C before use. The volumetric Karl Fischer method according to ASTM E203-16 is acceptable for confirmation when the moisture concentration exceeds 5 ppm, but at 1 ppm the coulometric method is preferred due to its lower detection limit and smaller sample demand. The online and laboratory data are compared in a monthly control chart with a bias allowance of 0.2 ppm; systematic deviations greater than this trigger a review of sample conditioning and reagent interference.

In a direct chlorination unit receiving ethylene from a naphtha cracker, splitter off-spec ethylene may contain acetylene and other unsaturates, but the moisture content of the ethylene after cold separator and dryers is normally stable below 0.1 ppmv. However, upstream trip events and dryer regeneration cycles create transient water pulses that can overwhelm a poorly designed ethylene feed gas filter coalescer. A coalescer element with a water removal rating of 0.3 µm is installed upstream of the molecular sieve dryer to protect the bed from free water slugs. The direct chlorination reactor also receives recycled EDC from the purification section; this recycle stream passes through a guard bed containing molecular sieve or a solid desiccant such as alumina to maintain water below 0.5 ppm. When the recycle guard bed reaches its water breakthrough point, the moisture in the reactor feed rises quickly from 0.2 ppm to 1 ppm and then to 2 ppm over a period of 2–6 h if the bed is not switched. The pressure drop across the guard bed rises as water absorption causes partial surface swelling of the desiccant material, and the bed outlet temperature may increase by 5–10°C due to the heat of adsorption. This measurable temperature front is used in some plants as an early warning of water breakthrough before the online moisture analyser trend begins to rise. Field operating procedures for a 150,000 t/y direct chlorination unit specify that the recycle guard bed is replaced or regenerated when the outlet moisture reaches 0.5 ppm or the differential pressure increases by 0.2 barg, whichever occurs first. This preventive threshold protects the reactor from reaching the 1 ppm operational fault line during normal capacity operation. The interaction between the upstream ethylene dryer, the recycle guard bed, and the online moisture analyser delay time creates a control system dynamic in which the operator cannot rely on a single instrument but must evaluate the combined trend of dryer outlet dew point, guard bed temperature rise, and liquid recycle moisture.