Sour Water Stripper Offgas Desulfurization with Zinc Oxide Sorbent

In petroleum refineries, sour water stripper (SWS) overheads produce a saturated offgas stream at 0.5–2.5 barg and 85–120°C, containing 1–10 mol% H₂S, 5–30 mol% NH₃, 2–15 mol% CO₂, and balance steam. This gas, when routed to a thermal oxidizer or Claus unit, demands H₂S removal to below 10 ppmv to meet flaring regulations (40 CFR 60.18) and to protect downstream amine units from fouling. Fixed-bed zinc oxide chemisorption — a non-regenerable, high-selectivity desulfurization technology — achieves outlet H₂S concentrations of <1 ppmv at gas hourly space velocities (GHSV) of 500–2000 h⁻¹ under dry gas conditions. The exothermic reaction ZnO + H₂S → ZnS + H₂O (ΔH°₂₉₈ = -76.5 kJ/mol) proceeds irreversibly at temperatures above 200°C and is widely deployed in refineries processing sour crudes with high nitrogen content, where amine-based acid gas removal generates ammonia-rich offgas streams that degrade conventional liquid scavengers. The ZnO bed service life, typically 6–24 months, is determined by the stoichiometric sulfur loading capacity of the sorbent, mass transfer zone length, and the presence of competing species that either poison the active surface or lead to pore plugging. This discussion examines the multidimensional factors governing ZnO sorbent performance, guided by published plant data and rigorous chemical engineering principles.

The Chemisorption Mechanism and Thermodynamic Constraints of ZnO–H₂S Systems

The reaction between ZnO and H₂S is a gas–solid non-catalytic heterogeneous process described by a shrinking core model with a distinct reaction front progressing from the pellet exterior inward. Kinetic studies on commercial ZnO sorbents ( 3–5 mm cylindrical extrudates, 20–40 m²/g BET surface area, 0.2–0.4 cm³/g pore volume) indicate that the intrinsic rate constant, k₀, follows an Arrhenius expression with an activation energy of 55 ± 6 kJ/mol, as reported by Westmoreland and Ray (1977) in thermogravimetric analysis of ZnO single crystals. At 300°C, the equilibrium constant for ZnO sulfidation, expressed as log Kₑq = 5.2 (Kₑq = [H₂O]/[H₂S]), guarantees complete conversion of H₂S to ZnS provided the water vapor partial pressure does not exceed thermodynamic reversal thresholds. In practice, when steam content in the SWS offgas exceeds 40 vol%, the equilibrium begins to shift, reducing the maximum achievable sulfur loading by 15–30% compared to dry feed conditions. The reaction rate is governed by two consecutive resistances: external film mass transfer of H₂S from bulk gas to pellet surface, and intraparticle diffusion through the growing ZnS product layer. At temperatures below 250°C, intraparticle diffusion becomes rate-limiting, leading to an unacceptably shallow mass transfer zone (MTZ) that can exceed 30% of the total bed length, hastening breakthrough. Above 400°C, ZnO crystallite sintering accelerates, reducing surface area by more than 50% over 500 hours and degrading crush strength from an initial 4 MPa to below 1 MPa, as measured via ASTM D4179 for single pellet crush strength. Thermax and Haldor Topsøe technical bulletins recommend a conservative operating window of 280–380°C to balance kinetics and mechanical durability, with a maximum allowable bed pressure drop of 0.3 bar to prevent pellet fragmentation.

How Does the Presence of CO₂ and NH₃ Alter Bed Performance and Sulfur Uptake?

Sour water stripper offgas invariably carries ammonia and carbon dioxide, which introduce side reactions that challenge ZnO bed integrity. In the temperature range 120–200°C, gaseous NH₃ and CO₂ combine to form ammonium carbamate (NH₄CO₂NH₂) and ammonium carbonate ((NH₄)₂CO₃), both of which have sublimation points below 60°C but can condense as waxy solids on cooler piping walls and at the top of the sorbent bed if the gas temperature drops below the ammonium salt dew point. Dew point prediction using the DIPPR 801 database indicates that for a gas containing 15 mol% NH₃ and 10 mol% CO₂ at 1.2 barg, salt formation initiates at 135°C. Consequently, SWS offgas is customarily passed through a heat-traced coalescer and a preheater to maintain a minimum bed inlet temperature of 160°C, with electric tracing of the reactor shell set to 150°C. Ammonia itself does not react with ZnO but competitively adsorbs on acidic surface sites, reducing available active area for H₂S chemisorption. BET measurements on spent sorbents exposed to NH₃-rich feeds show a 10–15% reduction in micropore volume (< 2 nm) due to ammonium bicarbonate deposition. CO₂, on the other hand, reacts with ZnO at temperatures above 300°C to form zinc carbonate (ZnCO₃), which decomposes above 380°C to regenerate ZnO. This reversible cycle leads to periodic pellet swelling and crack propagation, lowering the pellet crush strength by an additional 20–30% over a typical 12-month run. To mitigate these effects, some operators install a pre-bed guard layer of low-cost calcium oxide pellets ( 6 mm, 90% CaO) that scavenges CO₂ via carbonation, extending the ZnO bed life by 18–25% as documented in a 2018 case study at a US Gulf Coast refinery processing 200,000 bpd of Maya crude. The guard bed also acts as a distribution layer, ensuring plug flow and minimizing channeling at the ZnO interface.

When Steam-to-Gas Ratio Exceeds 40%: Hydrothermal Stability and Pellet Integrity

The SWS overhead gas is saturated with steam at its operating pressure; after cooling and condensate knockout, the residual steam content typically ranges from 25 to 45 mol%. When the steam-to-dry gas ratio exceeds 0.67 (equivalent to 40 mol% H₂O), hydrothermal degradation of the ZnO sorbent matrix becomes a dominant deactivation mechanism. In the presence of high-temperature steam, the ZnO crystallites undergo Ostwald ripening, where smaller crystallites dissolve and redeposit on larger ones, drastically reducing the specific surface area from 35 m²/g to less than 10 m²/g within 1000 hours of continuous operation at 380°C, as shown by long-duration steam-ageing tests conducted according to DIN 51726. The loss of surface area directly impairs the initial reaction rate and broadens the mass transfer zone, causing early H₂S breakthrough. Additionally, steam promotes hydrolysis of the silica or alumina binders commonly used in extruded ZnO pellets (5–10 wt% Al₂O₃ or SiO₂), weakening inter-particle bridges. Compressive strength measured via ASTM D7084 drops below the recommended minimum of 2.5 MPa, leading to fines generation and increased pressure drop. To offset hydrothermal effects, several commercial sorbents incorporate a secondary spinel phase, such as ZnAl₂O₄ or Zn₂SiO₄, which acts as a sintering barrier. For instance, PURASPEC 2250 (Johnson Matthey) formulations with 7 wt% alumina demonstrate 15% higher retained surface area after 1200 hours under 50 vol% steam compared to unmodified ZnO. In practice, steam content exceeding 45 mol% mandates a derating of the design sulfur loading by a factor of 0.75, as per licensor guidelines from Haldor Topsøe for their HTZ-5 sorbent. Pre-cooling the gas to 40°C to condense and remove excess water, followed by reheating to 200°C, is a common engineering strategy to maintain a steady steam partial pressure below the threshold, albeit at the expense of increased furnace duty and pressure drop in the knock-out drum.

Optimal fixed-bed reactor design for ZnO desulfurization of SWS offgas demands precise attention to pellet loading, gas distribution, and pressure drop minimization. Axial-flow cylindrical vessels (L/D ratio 2.5–4.0) are loaded via sock or dense-loading methods to achieve a packing density of 0.95–1.10 g/cm³, with a void fraction of 0.38–0.42. Pellet size distribution is tightly controlled to 3–5 mm diameter, with a maximum allowable length-to-diameter ratio of 3.0 to prevent bridging. The pressure drop across the bed is estimated using the Ergun equation: ΔP/L = 150μ(1-ε)²/(ε³dₚ²) u + 1.75ρ(1-ε)/(ε³dₚ) u², where dₚ is the equivalent pellet diameter, ε is void fraction, μ is gas viscosity, and u is superficial velocity. For a typical SWS offgas flow of 5000 Nm³/h at 300°C and 1.5 barg, with a bed diameter of 2.2 m, the pressure drop is maintained below 0.15 bar by limiting the gas hourly space velocity to 800 h⁻¹. Any sudden increase in pressure drop beyond 0.5 bar signals bed plugging due to either ammonium salt deposition or fines accumulation, triggering an automatic bed isolation and switch to a standby vessel. In lead-lag configurations, the lead bed is typically designed to treat 70% of the total sulfur load before the lag bed takes over, with online H₂S analyzers triggering bed changeover at 5 ppmv breakthrough. The mechanical design of the vessel must account for thermal expansion of the dense ceramic packing; a hold-down screen with a dead weight of 0.15 bar equivalent is installed to prevent bed fluidization during pressure surges.

Sorbent Parameter	High-Surface-Area Type	High-Strength Stabilized Type	Test Standard
Shape / Nominal Diameter	Cylindrical extrudate 3.0 mm	Cylindrical extrudate 4.5 mm	—
BET Surface Area	35–40 m²/g	25–30 m²/g	ISO 9277
Pore Volume (Hg intrusion)	0.32–0.38 cm³/g	0.22–0.28 cm³/g	ISO 15901-1
Crush Strength (single pellet)	3.5–4.5 MPa	5.0–6.5 MPa	ASTM D4179
Sulfur Loading at Breakthrough (dry, 300°C)	20–24 wt% S	18–21 wt% S	ASTM D4468 (total S)
Attrition Loss (air jet, 1 h)	<1.5 wt%	<0.8 wt%	ASTM D5757

Guard Bed Configuration and Multistage Contactor Strategies

For SWS offgas streams with high fluctuating H₂S concentrations (5–10 mol%), single fixed-bed ZnO units become economically unattractive due to rapid bed saturation and frequent changeouts. A multistage contacting approach, incorporating a disposable iron oxide guard bed or regenerable activated carbon pre-bed, significantly improves overall sulfur management. The guard bed, operated at 200–250°C and 1.0 barg, removes up to 80% of inlet H₂S, reducing the load on the downstream ZnO polisher, which then achieves a final H₂S slip of <0.1 ppmv. Iron oxide sorbents (Fe₂O₃ content 85–90 wt%, bulk density 0.9 g/cm³) offer a sulfur loading of 0.3–0.5 kg S/kg sorbent, but their reaction rate is 2–3 times slower than ZnO; thus, they are sized for a GHSV of 300–600 h⁻¹. The combined guard-ZnO system extends the ZnO change-out interval by a factor of 4–6, as reported in a 2016 operating report from a Middle Eastern gas oil separation plant processing 18 MMscfd of SWS offgas. In such a cascade, pressure drop across each bed is monitored via differential pressure transmitters with 4–20 mA output, and inter-bed cooling may be needed to manage the adiabatic temperature rise of 20–40°C across the ZnO bed due to the exothermic sulfidation. Adsorbent selection is guided by the equilibrium-limited capacity of the guard material under humid conditions; for instance, iron oxide performance degrades by 30% when steam exceeds 30 mol%, necessitating upstream water knockout before the guard bed. The spent guard material, classified as non-hazardous in most jurisdictions, can be landfilled after passivation, whereas sulfided ZnO containing zinc sulfide requires stabilization to prevent leaching as per EU Council Decision 2003/33/EC limits for Zn in leachate (<50 mg/L).

Breakthrough Monitoring with Tunable Diode Laser Absorption Spectroscopy

The extremely low H₂S outlet target of <1 ppmv demands continuous, high-accuracy monitoring to trigger bed change-out before environmental exceedances occur. Tunable diode laser absorption spectroscopy (TDLAS) has become the industry standard for this service due to its fast response time (1 second) and ability to measure H₂S in the presence of high moisture without cross-interference from NH₃ or CO₂. Analyzers conforming to ASTM D7165 are installed on the outlet line with sample cells heated to 150°C to prevent condensation. Calibration is performed using certified H₂S/N₂ gas mixtures traceable to NIST SRM 1694a (1 ppmv H₂S). The TDLAS system measures the absorption peak at 1.578 µm, where H₂S exhibits a strong rotational-vibrational line free from CO and H₂O interference. The measurement validation frequency is set to 48-hourly zero and span checks, per EPA Method 15A. A duplicate online analyzer with a time-delay triggered by a mass flow controller allows automatic switchover upon deviation of >2% of full scale. In addition to TDLAS, total sulfur in the outlet gas is tracked via a dry colorimetric tape method conforming to ASTM D4468 on a weekly grab sample basis to cross-validate the online readings and detect any carbonyl sulfide (COS) slip, which ZnO beds do not fully hydrolyze at temperatures below 350°C. At 350–400°C, ZnO can hydrolyze COS to H₂S, which is then fixed, but the reaction proceeds at 20–40% of the H₂S reaction rate, so COS slip remains a potential compliance risk if the bed operates at the lower end of the temperature spectrum.

Parameter	Light Sour Crude SWS	Heavy Sour Crude SWS	Coker SWS Offgas
Inlet H₂S Concentration	2.0–3.5 mol%	5.0–7.5 mol%	8.0–12.0 mol%
Steam Content (after KO drum)	28–33 mol%	38–45 mol%	42–50 mol%
NH₃ / CO₂ Concentration	8% / 3% (typical)	18% / 10% (typical)	25% / 12% (typical)
ZnO Bed Inlet Temperature	290–310°C	320–350°C	340–380°C
Design GHSV	1000–1200 h⁻¹	700–900 h⁻¹	500–700 h⁻¹
Outlet H₂S at Breakthrough (TDLAS)	<0.5 ppmv	<1.0 ppmv	<1.5 ppmv
Typical Bed Life (single stage ZnO)	15–20 months	10–14 months	6–9 months
Average Sulfur Loading Achieved	21–23 wt%	18–20 wt%	15–18 wt%

Spent ZnO sorbent, containing 15–25 wt% sulfur as ZnS, is classified as a non-hazardous solid waste under Basel Code Y23 only after a rigorous stabilization procedure involving controlled oxidation with 0.5–1.0 vol% O₂ in N₂ at 300°C to convert acid-soluble ZnS to stable ZnO and elemental sulfur, with offgas H₂S concentrations continuously monitored to stay below 10 ppmv via a scrubber blower.