Refineries Face a “Sweetening” Split: Caustic’s Waste Bill vs. Adsorption’s Clean Polish

For decades, refineries have “sweetened” fuels with caustic soda to rip out hydrogen sulfide (H₂S), carbon dioxide (CO₂), and mercaptans (sulfur‑containing thiols). It works—and fast. But there’s a catch: the waste. Operators report disposal alone can cost millions of USD per year at scale (gasprocessingnews.com).

Now a contrasting playbook—adsorption using molecular sieves or activated carbon—promises ultra‑deep polishing, sub‑ppm clean‑up, and far less liquid effluent. The tradeoff is upfront spend and energy for regeneration, a shift that dovetails with tightening sulfur specs and wastewater rules (sse.co.th).

Chemical sweetening: caustic extraction

Refineries traditionally use caustic (NaOH) scrubbing to remove acidic contaminants—H₂S, CO₂, and mercaptans—from liquid fractions (gasprocessingnews.com) (gasprocessingnews.com). Fractional caustic extraction of light LPG/naphtha cuts mercaptans to less than 5 wt ppm sulfur—often much less than 1 ppm—in product gas (gasprocessingnews.com). Caustic also irreversibly removes CO₂ and H₂S, a benefit when lean amine treating is insufficient.

The downside is the spent liquor: a highly alkaline waste (pH>12) containing roughly 2–15% sodium and about 2–3% sulfides/mercaptides (mdpi.com). Disposal options—deep‑well injection, incineration, wet‑air oxidation—are costly or constrained. Wet‑air oxidation of sulfidic caustic can remove 50–95% of COD (chemical oxygen demand), but reactor fouling and high oxygen/catalyst costs make it expensive (gasprocessingnews.com). A special waste‑hauler service may cost millions of USD per year for a large plant (gasprocessingnews.com). In some cases, nearby pulp mills can co‑use spent caustic, but such synergies are rare. The net: low capital cost up front, but high operating cost from NaOH makeup and disposal liabilities.

Chemical sweetening requires precise handling of reagents; accurate chemical dosing is integral to these circuits, where equipment such as dosing pumps support controlled NaOH addition.

Mercaptan oxidation (MEROX) unit performance

The MEROX (mercaptan oxidation) unit is a catalytic extension of caustic sweetening. In a MEROX bed—usually an alumina‑ or carbon‑supported catalyst in mild caustic—dissolved oxygen oxidizes R–SH to R–S–S–R (disulfide oil). It effectively “sweetens” naphtha or LPG by converting mercaptans without hydrogenation; disulfides are nonvolatile and less corrosive (total sulfur is unchanged, but R–SH is gone) (gasprocessingnews.com).

Typical MEROX performance yields product mercaptan sulfur below 5 wt ppm S in treated naphtha and below 20 ppm in jet/kerosene (gasprocessingnews.com). It requires minimal NaOH (just to dissolve O₂) and creates only a relatively benign disulfide byproduct (often sold or recycled). The capital cost is significant (fixed‑bed reactors, catalyst), but once online the operating cost is low—essentially air compression and occasional catalyst replacement. Compared with once‑through caustic, catalytically regenerated caustic cuts NaOH consumption and waste dramatically; UOP reports a “vast reduction in caustic consumption and disposal” at the expense of higher CAPEX (gasprocessingnews.com).

Molecular sieves: dry, selective removal

Molecular sieves (zeolites—microporous aluminosilicates) physically adsorb polar or size‑selected molecules. In refineries they’re mainly used to dehydrate and capture trace contaminants. 3A and 4A sieves dry NGL or LPG streams to less than 1 ppm H₂O (about −100 °F dewpoint) for cryogenic units. Some zeolites (such as 13X or specialized Cu/Zn‑zeolites) can chemisorb H₂S or CO₂ at moderate temperature, enabling “dry” removal in olefins/alkylation circuits. Beds regenerate by pressure or high‑heat purging, so energy dominates OPEX. Effectiveness is very high—sieves can achieve sub‑ppm levels of targeted species; an industry write‑up notes “molecular sieves remove sulfur compounds for cleaner combustion” to meet stringent fuel specs (sse.co.th).

Drawbacks are cost and sensitivity: special vessels and instrumentation are required, and periodic regeneration around 250–350 °C makes OPEX intensive. Sieves are easily poisoned by heavy hydrocarbons or particulates, so feeds usually pre‑filter. Waste generation is minimal (off‑gas CO₂ or captured moisture); thanks to in‑situ regeneration, spent sieves are rarely discarded.

Activated carbon polishing and catalyst protection

Activated carbon (AC) beds remove organic and oxidizable compounds via adsorption. In refineries AC is used for polishing and specialty tasks: removing residual VOCs, trace H₂S/mercaptans, hydrogen cyanide from tail gas, or decolorizing fuels. Carbon can be impregnated with metals (Cu, Zn, Ag, etc.) to chemisorb sulfur species (e.g., Cu‑AC for H₂S). With extremely high surface area (about 1000–1500 m²/g), AC captures trace‑level pollutants at ppb–ppm. A sulfoxide‑impregnated carbon claims “removal accuracy <0.05 ppm” mercaptans (colincarbon.com), and case studies show AC can extend catalyst life: a U.S. refinery saw a 15% drop in hydrotreater catalyst changes, saving about $2.8 million per year after installing AC guards (pmarketresearch.com).

There are tradeoffs. Once “spent,” carbon must be regenerated (steam or thermal) or replaced. Thermal regeneration consumes fuel and yields off‑gases (scrubbing or incineration required), so AC produces solid waste (spent carbon) and CO₂. On the upside, regeneration is viable; a demo showed thermal reactivation cut fresh carbon purchases by about 40%, slashing both OPEX and waste (pmarketresearch.com). Carbon beds also carry higher pressure drop and occasional dust issues. Compared to liquid caustic, AC generates no corrosive effluent—the principal waste is used carbon, often classed as hazardous due to adsorbed toxins. In refinery service, activated carbon is a modular, retrofit‑friendly way to polish streams and protect downstream catalysts.

Effectiveness, cost, and waste comparison

Effectiveness: Caustic plus MEROX handles bulk sour gas with very high removal—more than 99% of H₂S and mercaptans driven to single‑ppm levels (gasprocessingnews.com) (gasprocessingnews.com). Adsorption targets trace contaminants at sub‑ppm; a molecular sieve or Cu‑AC adsorber can polish fuel to tens of ppb sulfur if needed (colincarbon.com) (sse.co.th). In practice, refineries often combine both: caustic/MEROX for bulk removal, then sieve/AC polishing to hit ultra‑low specs.

Cost: Caustic sweetening has low CAPEX but high chemical expense and disposal cost (gasprocessingnews.com) (gasprocessingnews.com). One telex: fresh caustic can cost millions (USD/yr) for a large refinery, plus disposal costs of similar magnitude (gasprocessingnews.com). MEROX units cost more upfront (catalyst, fixed beds) but dramatically cut NaOH use (gasprocessingnews.com). Zeolite beds have the highest CAPEX (heater, control system), whereas activated carbon units sit in between. Operating OPEX for adsorption is moderate (mostly energy for regeneration or compressed air), with no continuous purchase of commodity chemicals. Adsorbents don’t create liquid waste, saving on effluent treatment.

Waste and environmental: Spent caustic is the dominant waste stream for caustic sweeper circuits; it has very high COD/sulfur load and must be oxidized or neutralized. Regulatory pressure on VOCs, SOₓ, and wastewater quality is tightening globally. In Indonesia, future Euro‑4/5 fuel rules (50 ppm sulfur) motivate ultra‑deep removal (transportpolicy.net). Adsorption avoids caustic sludges: used carbon/sieve can be thermally regenerated many times; the main waste is secondary (spent carbon and regeneration/incineration CO₂).

Typical figures and market notes

GasProcessing News (Honeywell) reports that well‑designed caustic extraction can reduce mercaptans to less than 1 wt ppm S and remove essentially all H₂S, though bed design and fractionation are critical (gasprocessingnews.com). In one survey, caustic regenerators (half‑MEROX) achieved more than 95% reduction of mercaptans with caustic loss low. Activated carbon, by comparison, has adsorption capacities on the order of 50–200 mg S per g of carbon for sulfides; vendors claim ppm‑to‑ppb‑level polishing. The AC market in refining is growing: global regulations (IMO 2020, Tier 3) have made carbon adsorbers commonplace (pmarketresearch.com). One refinery investment analysis credits AC with more than $2.8M/yr in savings from fewer catalyst changes (pmarketresearch.com).

Pros and cons at a glance

• Caustic washing: + Very effective for acid gases; low equipment cost. – Large waste streams (spent caustic, emulsions), ongoing NaOH cost, corrosion.
• MEROX: + Converts mercaptans efficiently with little chemical, minimal “wet” waste. – Significant CAPEX, not applicable to H₂S, total sulfur unchanged (disulfides remain).
• Molecular sieve: + Ultra‑deep removal (water, sulfur, VOCs) to ppb levels; cyclic/regenerable, no liquid effluent. – High capital and energy for regeneration; feed must be clean (no heavy oils); catalyst poisons if unprotected.
• Activated carbon: + Broad‑spectrum adsorption (sulfur, organics, metals); modular and retrofit‑friendly; moderate cost and pressure drop. – Periodic regeneration/disposal needed; less selective; lower throughput for high‑concentration loads.

Bottom line and sources

Caustic/MEROX remains a workhorse for bulk sweetening, but with heavy waste costs (gasprocessingnews.com). Adsorbers (sieves/AC) offer greener polishing to meet tight specs (transportpolicy.net) (sse.co.th), at the price of higher upfront investment and energy use. Optimal design often combines both: use chemical treaters to remove the bulk of H₂S/mercaptans, then polish with adsorbents to achieve ultra‑low sulfur and purity targets with minimal waste.

Authoritative sources back the metrics and tradeoffs: UOP/Honeywell engineers note fractionated caustic extraction can achieve less than 5 wppm*Sª in light fractions (even less than 1 wppm*S in some cases) (gasprocessingnews.com), but produces spent caustic (2–3% HS⁻) at pH>12 (mdpi.com). Waste‑hauling can “exceed millions of dollars per year” (gasprocessingnews.com). Activated carbon adsorbents reduce catalyst fouling—one refinery saw a 15% drop in HDS catalyst replacement, saving about $2.8M/yr with carbon beds (pmarketresearch.com). Regulatory trends (IMO 2020, Tier 3, Euro standards) are driving wider adoption of adsorption polishing (pmarketresearch.com) (transportpolicy.net). Both chemical and adsorption routes demand corrosion‑resistant materials and generate off‑gases (CO₂, vents); MEROX has minimal sludge but yields “spent air” (CO₂, minor organics) and a raffinate rich in NaHS that must be neutralized.