Why Delayed Coking Sour Water Reads High COD: Two Contributors, Two Different Fixes

The Reading That Starts the Investigation

Delayed coking sour water leaves the stripper with a COD that does not match the visible oil. Operators see a reasonably clear stream and a number that says otherwise. The gap is not a measurement error. It comes from two contributors that behave very differently, and treating them as one problem is why the fixes usually disappoint.

Contributor One: Oil That Will Not Coalesce

Where the surfactant system comes from

Refineries blending heavy, low-quality crude bring in large amounts of natural surfactants: asphaltenes, resins, petroleum acids. Add oil-production additives such as alkylbenzene sulfonic acid and corrosion inhibitors such as lye, and the sour water ends up carrying an anionic surfactant system that nobody formulated on purpose.

What that does at temperature

At high temperature the anionic surfactants lower the oil-water interfacial tension sharply. A strongly adsorbed layer forms on the droplet surface, together with a thick Stern electric double layer, and the result is a high zeta potential. Repulsion between droplets rises, a firm interfacial film forms, and coalescence and settling are blocked. Dissolved oil and emulsified oil content therefore climb, and COD climbs with them. The more heavy crude in the blend, the more pronounced the effect, which is why the problem appears after a crude slate change rather than gradually.

Contributor Two: Additives That Do Not Biodegrade

The secondary source is coke tower antifoaming agents, which contain alcohols, fatty acids and polyether compounds. They enter the water when the foam layer runs too high, when the antifoaming agent is over-injected, or when gas velocity inside the tower is too great and carryover reaches the fractionator overhead condensate. These compounds resist biodegradation and consume oxygen as they degrade, pushing COD up. Their mass flow is limited though, so their share of the total is usually modest: enough to blur a mass balance, not enough to explain the number on its own.

Contributor Three: Phenols, a Smaller Share Than Expected

Phenolic organics carry a COD conversion ratio of about 2.38 and have large molecular weights, which makes them hard to strip. But volatile phenol content in coking sour water is generally low, so the aggregate contribution stays comparatively small. This is the contributor most often over-estimated in troubleshooting, and over-estimating it sends money into the wrong unit operation.

Contributor Four: Sulfides, and the Arithmetic Is Unforgiving

Sulfides, mainly H2S, HS- and S 2-, present in a ratio of about 1:6:3, are strongly reducing species. Their oxidation consumes a large amount of oxygen, and that consumption is counted directly in the measured COD value.

Quantitatively, oxidising each milligram of sulfide consumes about 1.6 mg of oxygen, so the contribution ratio of sulfide to COD is about 1:1.6. In one actual case the sulfide content of a stripper feed was 3520 mg/L, corresponding to a theoretical COD contribution of about 5632 mg/L. The measured total COD of the same sample was 9110 mg/L. The sulfide share is not a rounding term. It is the majority of the reading.

What This Means for the Fix

Because the two main contributors need different treatment, the sequence matters. Oil has to be removed physically, and the emulsified fraction is the part that needs help, since a conventional separator will not break a film stabilised by an anionic surfactant system at high zeta potential. Sulfide has to be converted or stripped rather than oxidised in place, because oxidising it in the water column spends oxygen and shows up as COD instead of being removed.

The consequence of getting the order wrong is visible downstream. Large-molecular-weight oils survive conventional stripping, so the purified water still carries high COD after the stripper, around 2500 mg/L in the case above. Membrane systems downstream then see organic loading they were not designed for, and the failure mode is accumulation followed by collapse rather than a gradual decline.

So the practical target list is short: efficient oil removal at the front, sulfide conversion or removal before the water reaches the membrane train, and a COD budget that keeps the two contributions apart instead of reporting one number.

How to Split the COD in Practice

Two measurements are enough to start. Sulfide by iodometric titration, converted with the 1:1.6 ratio, gives the reducing share. Oil and grease on the same sample, after solvent extraction, gives the organic share. Whatever is left is phenol, additives and the residual dissolved fraction, and its size tells you whether an advanced oxidation step is justified at all.

The other useful check is upstream rather than downstream. Sample the stripper feed and the stripper bottoms on the same day and compare the sulfide drop against the oxygen demand drop. When the two move together the stripper is doing its job. When the oxygen demand stays flat while sulfide falls, the residual COD is organic, and no amount of stripping will move it.