Spent Caustic Oxidation Boundary Control in Liquid Hydrocarbon Sweetening

The spent caustic stream generated in liquid hydrocarbon sweetening contains sodium mercaptides, sodium sulfide, polysulfides, and unreacted mercaptans, and its oxidation boundary is defined by the point at which mercaptide conversion to disulfide oil is complete without exceeding the oxygen input that initiates thiosulfate and sulfate formation. In continuous refining operations, this boundary moves with feedstock sulfur speciation, caustic concentration, air supply pressure, and reactor residence time, so a single setpoint cannot be applied across kerosene, jet fuel, LPG, and sour condensate services. The oxidation step is normally conducted after liquid-liquid contact between hydrocarbon and a sodium hydroxide solution that contains a sulfonated cobalt phthalocyanine catalyst; the spent caustic is then contacted with air in a separate oxidation vessel or in-line mixer. The term boundary control refers to the simultaneous management of oxygen-to-sulfur stoichiometry, temperature, caustic inventory, and phase separation so that disulfide oil separates as a low-density phase and the regenerated caustic remains free of excessive thiosulfate and sulfate. Published data for specific refinery configurations is limited, but the common design envelope for sweetening-derived spent caustic oxidation indicates that oxygen supply must be matched to the molar sum of mercaptan and sulfide species, not to total petroleum sulfur.

What Limits Dissolved Oxygen Transfer in Spent Caustic Oxidation Reactors?

The rate-limiting step in spent caustic oxidation is frequently the interfacial transfer of oxygen from dispersed air bubbles into the aqueous caustic phase, not the homogeneous oxidation kinetics of sodium mercaptide. In refinery vessels fitted with perforated-pipe spargers or static mixer air injection, the volumetric mass-transfer coefficient depends on bubble diameter, gas holdup, caustic viscosity, temperature, and the concentration of surface-active naphthenic acids carried into the caustic from the upstream hydrocarbon feed. Gas holdup values measured in pilot-scale air-sparged caustic columns typically range from 0.04 to 0.12, and the corresponding oxygen transfer efficiency declines when bubble coalescence is promoted by hydrocarbon contamination. The air discharge pressure must exceed the combined hydrostatic head, vapor pressure, and control valve pressure drop; for a reactor liquid level of 6 m, this requirement commonly places the compressor discharge between 0.35 MPa g and 0.70 MPa g, depending on the downstream oxidation vessel design pressure. A vertical oxidation tower with an L/D ratio of 3.5:1 to 5.0:1 and a gas disengagement zone of 1.2 m is often used because it allows the spent caustic to remain in contact with air for 30–60 min without excessive carryover of caustic droplets into the disulfide oil product. Field experience from air-sparged spent caustic oxidation vessels has shown that sparger hole plugging by iron sulfide and thiosulfate scale accelerates when the spent caustic feed contains more than 50 mg/L suspended solids, and this plugging changes bubble size distribution long before the air flow transmitter records a deviation. The operational response to sparger plugging is therefore not a simple air flow increase; raising the air flow through a partially blocked sparger increases the gas velocity through the remaining open holes, produces larger bubbles, reduces interfacial area, and can push more oxygen than necessary into the reactor after the scale releases, producing a temporary sulfate excursion that is difficult to reverse.

Oxygen solubility in sodium hydroxide solutions decreases as caustic strength and temperature increase. This physical constraint means that a spent caustic oxidation unit operating with 15 wt% NaOH at 55 °C has a lower dissolved oxygen ceiling than the same vessel operating at 45 °C, even if the air supply pressure is unchanged. The boundary between mercaptan-limited and oxygen-limited operation therefore shifts with caustic strength; a rise in caustic concentration from 10 wt% to 20 wt% may reduce the oxygen mass-transfer coefficient by a factor that field operators observe as longer disulfide oil breakthrough times and higher residual mercaptan in the regenerated caustic. In addition, surface-active naphthenates and phenolic compounds extracted from cracked stocks stabilize foam at the gas-liquid interface. The resulting foam layer in the oxidation vessel reduces the effective dense-phase height, lowers liquid residence time, and can carry disulfide oil into the off-gas system if the foam is not controlled by demister internals or by operating the vessel at a temperature high enough to reduce foam stability. The acceptable foaming boundary is commonly determined by differential pressure measurement across the vessel, with a rapid increase in top-bed differential pressure indicating that the air-to-caustic ratio has exceeded the local disengagement capacity.

The following ranges are representative of pilot-scale evaluations using sulfonated cobalt phthalocyanine catalyst and air-sparged oxidation vessels; they illustrate the shift from incomplete mercaptide conversion to sulfate overproduction as excess oxygen and temperature rise.

These data show that the upper oxygen boundary is not defined by complete mercaptide conversion, because conversion approaches 100 % before thiosulfate and sulfate generation reach their highest values. A control scheme based on mercaptan conversion alone would therefore permit excessive oxygen input. The practical control point is the inflection in thiosulfate and sulfate generation; for the conditions shown, this inflection occurs near 15–20 % excess oxygen. Above this band, sulfate formation accelerates and the disulfide oil separator must handle a larger rag layer of sodium sulfate solids and emulsified disulfide oil. The lower oxygen boundary is set by residual sodium sulfide in the regenerated caustic, since sodium sulfide reacts with dissolved iron in downstream equipment and can release hydrogen sulfide if the regenerated caustic is pH-adjusted or mixed with acidic streams.

In sour LPG prewash systems, the oxidation boundary is dominated by the molar ratio of hydrogen sulfide to mercaptan in the spent caustic. Straight-run propane and butane often contain 10–150 ppmv hydrogen sulfide and 20–500 ppmv methyl and ethyl mercaptans, while coker-derived LPG can contain 1,000–10,000 ppmv hydrogen sulfide and a much lower mercaptan concentration. The spent caustic from a prewash contactor operating at 1.2–1.8 MPa g and 35–50 °C therefore has a sulfide-to-mercaptide molar ratio that may vary from 0.5:1 to more than 3:1 within a single operating shift if the feed is changed from storage to coker product. This variation shifts the oxygen demand by more than 250 % and invalidates a fixed air-to-caustic ratio. A more stable control boundary is obtained by measuring the spent caustic oxidation-reduction potential and the sulfide-to-mercaptide ratio by ion chromatography, then cascading the signal to the air flow controller with a feed-forward multiplier based on spent caustic flow. Incomplete oxidation in high-sulfide service leaves sodium sulfide in the regenerated caustic; when this stream is recycled to the LPG contactor, the residual sulfide can reabsorb hydrogen sulfide more slowly than fresh caustic, but its more serious consequence is the precipitation of iron sulfide in the caustic circulation lines and the fouling of the air sparger. Field observations from LPG sweetening units with caustic circulation rates of 8–20 m³/h indicate that sparger fouling is most common when the oxidation temperature is below 45 °C, because thiosulfate and iron sulfide precipitation are enhanced at lower temperatures and higher caustic strengths.

When Caustic Strength Falls Below the Mercaptide Hydrolysis Threshold

The lower caustic concentration limit is not determined solely by extraction efficiency; it is also set by the hydrolysis equilibrium of sodium mercaptide. At sodium hydroxide concentrations below 5 wt%, methyl and ethyl mercaptans partition back into the regenerated caustic vapor space, and the off-gas from the oxidation vessel may exceed odor thresholds even when liquid-phase mercaptide conversion appears complete. The equilibrium RSH + NaOH ⇌ RSNa + H2O shifts to the left as caustic strength decreases, so a regenerated caustic stream that discharges from the oxidation vessel at 4 wt% NaOH may still contain hydrolyzed mercaptan in the aqueous phase. Conversely, caustic concentrations above 20 wt% increase kinematic viscosity and reduce the rate of disulfide oil coalescence, leading to rag-layer accumulation in the disulfide separator. The practical operating band for most extracted mercaptan systems is 10–20 wt% NaOH, with 15 wt% used as the design midpoint for jet fuel sweetening units. Sodium hydroxide concentration is not measured reliably by specific gravity alone when dissolved sodium thiosulfate and sodium sulfate are present; acid-base titration against phenolphthalein or automated pH titration is required. The oxidation boundary is also affected by water balance: oxidation of sodium mercaptide to disulfide oil generates sodium hydroxide and consumes water, while oxidation of sodium sulfide to thiosulfate also consumes water. If the spent caustic feed contains high sulfide and the air supply is controlled tightly, the regenerated caustic may actually increase in hydroxide concentration; if the air supply is pushed to sulfate formation, the caustic concentration may decrease because sodium sulfate formation consumes additional hydroxide.

The analytical compliance boundary for liquid hydrocarbon sweetening and spent caustic oxidation is anchored to the following standard test methods and operating limits.

Sampling frequency for each parameter is not fixed by the test method but by the variability of the feed sulfur species and the consequence of exceeding the boundary. For aviation turbine fuel, mercaptan sulfur by ASTM D3227 should be checked at least every 4 h during feed switches, because a single high-sulfur crude can raise the extractor mercaptan load above the oxidation unit design capacity. The Doctor test by ASTM D4952 is less quantitative but remains a rapid bulk verification for active sulfur in product tankage. In spent caustic service, the sodium hydroxide concentration should be measured once per shift at a minimum, but the oxidation air-to-sulfur ratio should be recalculated continuously from the spent caustic flow, sulfide, and mercaptide analytical data; manual calculation is insufficient when the feed contains variable cracked stocks.

Simultaneously, the oxidation boundary is sensitive to catalyst deactivation and metal precipitation, which are not detected by routine hydrocarbon product sulfur measurements. Sulfonated cobalt phthalocyanine catalyst is active in the caustic phase, but it can precipitate as cobalt hydroxide or adsorb onto iron sulfide solids when the pH drops or when dissolved iron is introduced from upstream piping. Iron at concentrations above 10 mg/L in the caustic stream is a known fouling precursor, and its presence is often accompanied by a loss of oxidation activity that is mistakenly attributed to low air flow. The resulting operator response is an air rate increase, which pushes the system across the upper oxidation boundary and generates sulfate scale without restoring catalyst activity. A better boundary parameter is the catalyst concentration in the caustic, measured by ultraviolet-visible absorbance at the characteristic wavelength of the sulfonated cobalt phthalocyanine complex; when the absorbance falls below a site-specific minimum, the caustic inventory should be purged and re-charged rather than over-oxidized. The purge boundary is also affected by the buildup of sodium thiosulfate and sodium sulfate, which cannot be removed by disulfide oil separation and which accumulate in a closed caustic loop. Published data for the maximum tolerable thiosulfate concentration varies with the downstream equipment metallurgy and disposal route, but many refiners limit thiosulfate sulfur to 5,000 mg/L in recycled caustic to avoid salting and to maintain adequate caustic strength.

Air-to-Sulfide Stoichiometry, ORP Trend Reversal, and Sulfate Scale Formation

The oxidation boundary for spent caustic carrying sodium sulfide is a two-stage oxygen demand. In the first stage, sodium sulfide is oxidized to sodium thiosulfate: 2 Na2S + 2 O2 + H2O → Na2S2O3 + 2 NaOH. In the second stage, thiosulfate is oxidized to sodium sulfate under excess dissolved oxygen and elevated temperature: Na2S2O3 + 2 O2 + 2 NaOH → 2 Na2SO4 + H2O. The second stage is the principal cause of sulfate scale in oxidation reactor effluent coolers and caustic preheaters, because sodium sulfate solubility drops sharply as temperature falls from 55 °C to 35 °C. A control strategy that permits high excess air during a transient mercaptan peak may therefore create a delayed sulfate precipitation event in the cooler downstream of the oxidation vessel, with the scale appearing only after the temperature has dropped and the sulfate concentration has exceeded the local saturation limit. Sulfate scale in the caustic preheater is particularly damaging because it reduces heat transfer and creates hot spots that accelerate thiosulfate oxidation. The lower oxygen boundary is marked by residual sodium sulfide in the regenerated caustic; the upper oxygen boundary is marked by a rise in sulfate concentration and a decrease in caustic pH due to hydroxide consumption in the sulfate-forming reaction. Between these boundaries, the oxidation-reduction potential signal from a platinum electrode often shows a characteristic plateau during mercaptide oxidation and a sharp rise when sulfide has been consumed and excess dissolved oxygen begins to accumulate. The exact millivolt reading varies with pH, temperature, and the reference electrode type, so the ORP instrument should be used as a trend indicator rather than an absolute setpoint unless it is calibrated against the specific caustic matrix.

Excess air is normally maintained at 10–20 % above stoichiometric oxygen demand. Values below 5 % cause mercaptan breakthrough and leave sulfide in the regenerated caustic; values above 25 % increase sulfate formation. The stoichiometric demand itself is calculated from the molar concentrations of methyl mercaptan, ethyl mercaptan, propyl mercaptan, and hydrogen sulfide in the spent caustic feed, using the oxidation reactions for mercaptide and sulfide. For a spent caustic stream containing 2,000 mg/L mercaptan sulfur and 800 mg/L sulfide sulfur, the oxygen demand is dominated by the sulfide fraction because sulfide consumes 2 mol oxygen per 2 mol sodium sulfide to form thiosulfate, while mercaptide consumes only 0.5 mol oxygen per 2 mol mercaptide. The air flow setpoint should be corrected for spent caustic flow, temperature, and pressure, and the analyzer delay time from the sample point to the ion chromatograph should be included in the feed-forward loop. A feed-forward-only control scheme without periodic feedback from residual sulfide or ORP will drift because the sulfide-to-mercaptide ratio changes with crude slate and upstream caustic prewash severity.

Regenerated Caustic Recycle Boundaries in Kerosene and Jet Fuel Sweetening Are Set by Thiosulfate Accumulation and Disulfide Oil Solubility

In kerosene and jet fuel treaters, regenerated caustic can be recycled to the extractor only when the oxidation step has converted sodium mercaptide to disulfide oil and the disulfide oil has been removed to a concentration below 500 mg/L as sulfur. Higher disulfide oil carryover increases the equilibrium sulfur content of the treated product because disulfide oil is partially soluble in the hydrocarbon stream and is not extracted by the caustic solution. The recycle boundary is therefore set by analyzer readings for mercaptan sulfur by ASTM D3227 on the treated fuel, not by visual caustic clarity. A caustic stream that appears clear can still contain 200–400 mg/L disulfide oil sulfur, and this material can raise the total sulfur of a jet fuel that otherwise meets the mercaptan specification. The disulfide oil separator must therefore be designed for a residence time of at least 30 min and a continuous skimming rate based on the disulfide oil generation rate; batch skimming is insufficient when the extractor feed mercaptan concentration varies by more than 100 mg/kg sulfur. Thiosulfate accumulation imposes a second recycle boundary because sodium thiosulfate cannot be separated by gravity and its concentration increases with each closed-loop cycle. In units where the caustic is not purged or slipstreamed to waste treatment, thiosulfate sulfur can reach 3,000–5,000 mg/L within several weeks, at which point the solution viscosity increases and the disulfide oil separation deteriorates. The practical recycle limit is therefore a combination of disulfide oil carryover, thiosulfate concentration, and caustic strength, and all three must be measured before the regenerated caustic is returned to the extractor.

Continuous oxidation boundary control requires an analyzer suite that includes ion chromatography for sulfate, thiosulfate, and sulfide, ultraviolet-visible spectroscopy for catalyst concentration, and an in-line ORP probe with automatic temperature compensation. The air flow control valve is positioned from a feed-forward calculation based on spent caustic flow, sulfide, and mercaptide data, with a feedback trim from the ORP or residual sulfide analyzer. When the spent caustic oxidation vessel is operated in a refinery with frequent crude switches, the analyzer sample line must be heat-traced to 50–60 °C to prevent sodium sulfate and thiosulfate crystallization in the sample tubing. The oxidation vessel temperature is controlled by the air preheater and by the exothermic oxidation reactions; a temperature rise above 60 °C is not necessarily required for mercaptide conversion and often indicates excessive oxygen input or reduced spent caustic flow. The lower temperature limit is set by disulfide oil viscosity and the risk of sodium sulfate scaling; operation below 40 °C is generally avoided in spent caustic oxidation units that process cracked stocks. The boundary between acceptable and unacceptable operation is therefore a multi-dimensional envelope that must be maintained with continuous analytical feedback, and any single-parameter control strategy will fail when the feedstock sulfur speciation changes.