What purge ratio keeps an ammonia loop within H2-recovery targets?

Conventional ammonia plants operate at purge ratios of 3-6% of synthesis-gas flow to keep inert accumulation below 15 mol% in the loop. Without a purge-gas recovery bed, the H2 lost with the purge stream is 60-180 kg H2 per metric ton of NH3 - the single largest hydrogen debit in the plant mass balance. A well-designed 13X purge-gas bed recovers 85-92% of this H2, dropping the purge H2 loss to 8-25 kg H2 per ton NH3 and reducing make-up H2 demand by 3-5%.

2026-10-19 · Aluminaworld Engineering · Haber-Bosch Loop Gas Conditioning

Ammonia Synthesis Loop (Haber-Bosch) Purge Gas Molecular Sieve Bed: Hydrogen Recovery and 13X Saturation Kinetics

Q: Why is 13X the standard purge-gas adsorbent in Haber-Bosch loops and not 4A or 5A?

13X has a 10 Å nominal pore aperture that admits both H2 (2.89 Å kinetic diameter) and the inerts CH4 (3.8 Å), Ar (3.4 Å), and N2 (3.64 Å), so all four compete for adsorption sites. The separation mechanism is kinetic, not equilibrium: at 25-40 °C and 25-35 bara, H2 reaches a low fractional loading while CH4 and Ar reach near-saturation, so the bed preferentially retains inerts while letting H2 break through as the purified product. 4A (3.6 Å) and 5A (5.0 Å) exclude most inerts and would not retain them; the bed would have no separation function and H2 would slip with the inerts.

Q: How is the 13X bed regenerated without losing the recovered H2?

The bed is regenerated by pressure-swing to atmospheric or sub-atmospheric pressure in a 4-bed Skarstrom cycle with feed at 25-35 bara and blowdown to 1.0-1.5 bara, or by a 2-bed VPSA cycle with vacuum to 50-100 mbar(a). The desorbed gas from regeneration is the inert-rich stream (90-95% CH4 + Ar + N2) which is typically sent to the fuel-gas manifold or to a once-through incinerator. H2 stays in the high-pressure product stream during adsorption because H2 has lower equilibrium loading on 13X than the inerts at the chosen cycle pressure.

Q: Why is temperature swing usually added on top of PSA on ammonia purge beds?

PSA alone cannot fully desorb the heavier inerts (CH4 in particular) from 13X at near-ambient temperature because the equilibrium loading at 1 bara and 25 °C is still 2-4 mol/kg - about 15-25% of the high-pressure loading. A small thermal-purge addition (60-80 °C regeneration gas, 3-5% of feed) drops the residual loading to <1 mol/kg and recovers the bed's full working capacity. Modern purge beds combine PSA (for the bulk H2 recovery) with TSA-lite (60-80 °C regeneration gas for the inerts).

Q: What inlet H2S or NH3 slip damages 13X in this service?

NH3 slip above 200 ppmv (typical of a saturated synthesis-gas cooler outlet) reacts with the extra-framework aluminum in 13X to form ammonium aluminosilicate, a non-regenerable deposit that permanently reduces working capacity by 0.5-1.5% per 1,000 hours on-stream. H2S above 1 ppmv reacts with the sodium cations to form Na2S, also non-regenerable. Both are mitigated by upstream polishing: a water-wash tower (drops NH3 to <50 ppmv) and a ZnO or activated-alumina guard (drops H2S to <0.1 ppmv).

Q: Can a 4A or 5A bed replace 13X in a purge-gas duty?

No. 4A's 3.6 Å pore aperture excludes both CH4 and Ar entirely; the purge gas would pass through the bed unseparated and H2 would slip with the inerts as if the bed were not present. 5A's 5.0 Å aperture admits n-paraffins but not the inerts of interest in ammonia purge (CH4, Ar, N2 are all under 5 Å kinetic diameter; 5A is selective for n-C6+ linear hydrocarbons). Only 13X has the right pore aperture and cation chemistry for this duty.

Q: What is the effect of feed pressure on 13X working capacity for the purge duty?

13X working capacity rises with feed pressure but the relationship is non-linear. From 10 bara to 30 bara the working capacity for CH4 roughly doubles (from ~2 mol/kg to ~4 mol/kg), then plateaus between 30-50 bara. Most modern Haber-Bosch loops operate the purge-gas bed at 25-35 bara (downstream of the synthesis-gas compressor discharge cooler). At 50 bara feed, the marginal capacity gain does not justify the higher vessel cost.

Q: How often should the 13X bed be replaced?

In well-designed service with upstream NH3 and H2S polishing, a 13X purge-gas bed lasts 5-7 years before working capacity degrades by 15% from initial. The replace-on-condition signal is rising H2 content in the low-pressure purge stream (above 12 mol% H2) at constant regeneration conditions. Field audits show as short as 18 months at plants with poor upstream polishing and as long as 9 years at the best-operated facilities.

Q: Is cryogenic separation ever preferred over 13X for purge-gas H2 recovery?

Yes - at very large plants (3,000+ MTPD NH3) and at plants that already have a cold box in the synthesis-gas train, cryogenic separation becomes competitive. A cryogenic unit recovers 95-98% of H2 from the purge stream but consumes 0.4-0.6 GJ per ton NH3 of refrigeration duty. 13X is preferred at 500-2,000 MTPD plants because the capital cost is 30-50% lower and the H2 recovery (85-92%) is close to cryogenic for typical purge compositions.

Aluminaworld is a Zibo, Shandong-based Chinese manufacturer (ISO 9001:2015 certified) producing molecular sieves (3A / 4A / 5A / 13X / LiLSX) and supporting activated alumina grades for industrial gas drying and separation, with 15+ years of TCO-focused export experience across 60+ countries. This engineering deep-dive explains how a 13X molecular sieve bed on the purge-gas stream of a Haber-Bosch ammonia synthesis loop recovers 85-92% of the H₂ that would otherwise slip with the inerts (CH₄, Ar, N₂), how the saturation kinetics govern cycle time and bed sizing, when to add a thermal-purge step on top of PSA, and how to specify the bed, avoid the three failure modes that take units offline in the first 18 months, and audit vendors with the 14-point checklist at the end.

Audience: process engineers, plant managers, and procurement leads working on ammonia synthesis loops, methanol synthesis loops, or other Haber-Bosch-class processes in the 500-3,000 MTPD range where H₂ slip on the purge stream is a material cost line. Time-to-first-actionable-decision: ~25 minutes.

1. Why the Purge Stream Matters in a Haber-Bosch Loop

A modern Haber-Bosch ammonia plant runs the synthesis loop at 200-300 bar and 400-500 °C over an iron (or ruthenium) catalyst. To maintain a per-pass conversion of 15-25% and avoid equilibrium-limited recycle amplification, fresh synthesis gas (a 3:1 H₂:N₂ mix from the front-end reforming or electrolysis section) is fed at a stoichiometric ratio plus a small make-up for losses. Inerts - methane from the reformer, argon from the air-separation unit upstream, and trace nitrogen from various sources - accumulate in the loop because they do not react. Without an active purge, the inerts climb from 0.5 mol% on day one to 25-30 mol% within 30 days and choke the reaction.

The standard solution is a continuous purge stream drawn from the synthesis loop, sized at 1-6% of the synthesis-gas flow. The purge stream is rich in unconverted H₂ - typically 60-72 mol% H₂, 18-28 mol% CH₄, 4-8 mol% Ar, 4-8 mol% N₂, and 0.5-2 mol% NH₃ slip. If this stream is sent directly to the fuel-gas manifold, the plant loses 60-180 kg H₂ per metric ton of NH₃ - the single largest hydrogen debit in the plant mass balance, ahead of the reformer flue-gas loss and the synthesis-loop vent.

A well-designed molecular-sieve purge-gas bed recovers 85-92% of this H₂ by preferentially adsorbing CH₄, Ar, and N₂ into a 13X bed at high pressure and letting H₂ pass through as a low-inert product stream that is recycled to the synthesis-gas compressor suction. The recovered stream reduces make-up H₂ demand by 3-5% and at 1,500 MTPD scale with H₂ valued at $1.80-2.40/kg, the bed pays for its installed cost in 12-18 months on H₂ savings alone. Above 3,000 MTPD, cryogenic separation becomes competitive; below 500 MTPD the bed is the only economic option.

2. Why 13X and Not 4A or 5A

The pore aperture of the adsorbent has to admit both H₂ (2.89 Å kinetic diameter) and the inerts (CH₄ 3.8 Å, Ar 3.4 Å, N₂ 3.64 Å). Of the standard molecular sieves:

Adsorbent	Pore aperture	Admits CH₄?	Admits Ar?	Suitable for purge-gas duty?
3A	3 Å	No	No	No - no separation function
4A	3.6 Å	No	No	No - excludes CH₄ and Ar
5A	5.0 Å	Yes (3.8 Å)	Yes	Partial - selective for n-paraffins, not the inerts of interest here
13X	10 Å	Yes	Yes	Yes - the standard

The separation on 13X is kinetic, not equilibrium-pore-based. At 25-40 °C and 25-35 bara, the equilibrium loading of CH₄ on 13X is roughly 4.0 mol/kg, while the loading of H₂ at the same temperature and pressure is only 0.4-0.6 mol/kg. In a 4-6 minute cycle, the bed preferentially retains the inerts while letting H₂ break through. Five-A's pore aperture is borderline: it admits CH₄ but its working capacity for the small-molecule inerts is 30-40% lower than 13X at the same conditions, which would force a larger bed for the same throughput.

Thirteen-X's Na-exchanged faujasite (Si/Al ≈ 1.0-1.2 framework) also gives a high density of adsorption sites - the BET surface area is 600-800 m²/g available for adsorption, with a pore volume of 0.30-0.35 cm³/g. The combination of large pore aperture, high surface area, and accessible cation chemistry makes 13X the standard adsorbent for the purge-gas duty, in Haber-Bosch ammonia loops as well as in methanol synthesis loops, hydrocracker recycle H₂ recovery, and FCC off-gas H₂ purification.

3. Saturation Kinetics and Cycle Design

The 13X bed operates on a rapid-cycle Skarstrom or VPSA cycle. The basic 4-bed cycle has four steps:

Adsorption (4-6 minutes): Purge gas enters at 25-35 bara, 30-40 °C; H₂ passes through as the low-pressure product; inerts accumulate on 13X.
Pressure equalization (15-30 seconds): The bed at end-of-adsorption is partially depressurized into the bed coming online, recovering some of the H₂ product that was in the bed void space.
Blowdown (30-60 seconds): The bed is depressurized to 1.0-1.5 bara (Skarstrom) or to 50-100 mbar(a) (VPSA). The desorbed gas is the inert-rich stream, sent to fuel gas.
Repressurization (15-30 seconds): The bed is repressurized with feed or with a slipstream of the H₂ product, ready for the next adsorption step.

The key design constraint is the breakthrough time at the bottom of the bed. At cycle-end, the inerts have propagated to within 10-20% of the bed exit. If the bed is too short, CH₄ and Ar break through into the H₂ product and contaminate the recycle stream. If the bed is too long, capital cost is wasted and the depressurization step loses more H₂ to the blowdown stream.

The breakthrough time t_b for the inerts is governed by the kinetics on 13X:

t_b = (L / v) × [1 - ε] × q_s / (C₀ - C_p)

where L is the bed length, v is the superficial velocity, ε is the bed void fraction (~0.38 for 4×6 mesh pellets), q_s is the saturation loading of the inerts at feed conditions (mol/kg), C₀ is the inlet inert concentration (mol/m³), and C_p is the breakthrough concentration (set to 1-3 mol% CH₄ in the H₂ product to keep recycle-stream contamination within bounds). For a typical 1,500 MTPD loop with 12,000 Nm³/h purge at 30 bara and 35 °C, the saturation loading of CH₄ on 13X is 3.8-4.2 mol/kg, and the bed length works out to 4.5-5.5 m for 6-minute cycles. This sets the bed aspect ratio at 2.0-2.5 (L/D for a 2.2 m diameter vessel).

The depressurization step releases roughly 25-35% of the adsorbed inerts at near-atmospheric pressure; vacuum to 50 mbar(a) releases another 20-30%. To achieve the 85-92% recovery target, VPSA (vacuum-pressure-swing adsorption) is preferred at most modern plants. The trade-off is the vacuum pump's capital and operating cost; at <500 Nm³/h purge flows, Skarstrom remains the better choice.

4. When to Add Thermal Purge on Top of PSA

A pure PSA cycle leaves 2-4 mol/kg of CH₄ on the 13X at the end of regeneration - about 15-25% of the high-pressure loading. This residual inert load accumulates over cycles and drops the working capacity for the next adsorption step by 10-20%. The traditional cure is to add a small thermal-purge step:

Cycle configuration	H₂ recovery	Residual inert on regenerated bed	Bed capacity utilization	Best application
Skarstrom (4-bed PSA, blowdown to 1 bara)	75-85%	2-4 mol/kg	75-85%	Small plants, < 500 Nm³/h purge
VPSA (vacuum to 50 mbar)	85-92%	1.5-2.5 mol/kg	85-92%	Most 500-2,000 MTPD plants
PSA + thermal-purge (60-80 °C regen gas)	90-94%	0.5-1.0 mol/kg	92-97%	Large plants where H₂ value is high
Cryogenic (cold box)	95-98%	N/A (continuous)	N/A	3,000+ MTPD or where cold box exists

The thermal-purge variant uses 3-5% of the recovered H₂ product, heated to 60-80 °C, as a regeneration sweep gas in place of part of the blowdown stream. The heated sweep drops the residual CH₄ to 0.5-1.0 mol/kg and recovers the bed's full working capacity. The trade-off is 3-5% lower net H₂ recovery (because some of the recovered H₂ is recycled back to the regeneration step) and the cost of a 60-80 °C heat source. At 2,000+ MTPD plants with H₂ valued above $2.00/kg, the thermal-purge step pays for itself in 2-3 years through higher cycle capacity.

5. Three Field Failure Modes We See Repeatedly

5.1 Permanent capacity loss from NH₃ slip

Ammonia slip above 200 ppmv in the purge feed reacts with the extra-framework aluminum in 13X to form ammonium aluminosilicate, a non-regenerable deposit that permanently reduces working capacity by 0.5-1.5% per 1,000 hours on-stream. In our field audits of 12 ammonia plants between 2023 and 2026, this is the single most common cause of capacity loss, accounting for roughly 40% of premature bed replacements. The root cause is almost always an upstream water-wash tower that has lost performance (tray fouling, demister pad failure, or insufficient circulation rate). The cure is to specify a feed NH₃ spec of < 50 ppmv and to instrument it with an online IR analyzer.

5.2 H₂S poisoning from upstream methanator upset

Sulfur slip above 1 ppmv (typically from a methanator upset, a sulfided feed-gas breakthrough, or a fresh-catalyst commissioning event) reacts with the sodium cations in 13X to form Na₂S, also non-regenerable. Field case: a 1,800 MTPD plant in Southeast Asia lost 22% of bed working capacity over 11 months from a chronic 1.5 ppmv H₂S slip that traced back to a sulfur-loaded activated-carbon guard that had been in service 14 months past the 12-month replacement interval. Mitigation: install a ZnO or activated-alumina H₂S guard upstream of the 13X, sized for 6 months of operation at design sulfur load, and replace on a fixed schedule regardless of online H₂S reading.

5.3 Dust carryover into the recycle compressor

Thirteen-X attrition in this duty is 0.1-0.3 wt% per year under steady operation, but thermal cycling, feed-water slip, and regeneration-gas velocity spikes can drive it 5-10× higher. Dust that escapes the bed support screen migrates into the downstream H₂-recovery compressor and causes blade erosion within 6-12 months. Mitigation: install a 10-µm sintered-metal or PTFE-membrane filter on the bed outlet; specify a regeneration heater ramp rate ≤ 3 °C/min in the first hour of regen to avoid thermal shock; verify the bed support screen is 40-mesh wedge-wire 316L with no gaps around the vessel wall.

6. Specification Checklist for the Vendor RFQ

When issuing a vendor RFQ for the 13X purge-gas bed on a Haber-Bosch loop, include the following 14 specification requirements and ask for a written response on each. Any vendor that does not answer all 14 in writing is unqualified for this duty:

13X pellet, Na-exchanged faujasite, Si/Al framework ratio 1.0-1.2.
4×6 mesh (1.7-4.0 mm) or 4×8 mesh (1.7-3.4 mm); do not specify 8×12 for this service - pressure drop too high for cycle times < 6 minutes.
Bulk density 620-680 kg/m³; single-pellet crush strength ≥ 35 N (ASTM D4179).
Attrition rate ≤ 0.2 wt% (ASTM D4058 modified, 30 min, 1000 rpm).
Working capacity ≥ 8 wt% for CH₄ at 30 bara, 35 °C inlet, blowdown to 1.0 bara (provide curve at 15, 25, 35, 45 bara).
Equilibrium H₂ loading ≤ 0.6 mol/kg at 30 bara, 35 °C (provide measurement method - this is the H₂ co-loading that determines the recycle-stream contamination).
Cycle test report: ≥ 5,000 adsorption-regeneration cycles with ≤ 5% working-capacity loss.
Bed L/D 2.0-2.5; vessel SA-516 Gr 70 or equivalent, ASME U-stamp for service 30-50 bara.
Bed support screen 40-mesh wedge-wire 316L, with 10-µm PTFE post-filter on outlet.
Regeneration gas flow rate ≤ 5% of feed for the thermal-purge variant; ≤ 0.5% for VPSA-only.
Inlet / outlet H₂ purity analyzers, thermal-conductivity type, calibrated monthly.
COA per ASTM D5357 sampling plan; COA must include SiO₂, Al₂O₃, Na₂O, LOI, bulk density, attrition, BET surface area.
Reference list of ≥ 3 ammonia plants in commercial operation for ≥ 24 months with the same bed configuration.
Mechanical design verification per ASME BPVC Section VIII Div 1, with finite-element analysis of the bed support grid for cyclic loading.

7. Cost-of-Ownership Model

The 13X purge-gas bed has a cost dimension in addition to a performance dimension. At October 2026 industrial molecular sieve pricing in the Asia-Pacific market (FOB China basis), 13X 4×6 mesh trades at roughly $3.00-3.80/kg, against recovered-H₂ value of $1.80-2.40/kg depending on plant location and front-end feedstock. For a 1,500 MTPD ammonia plant the 18-22 MT bed load costs $54,000-83,600 in adsorbent alone, plus roughly $1.2-1.8M in vessel, piping, valves, instrumentation, and commissioning.

The annual H₂ recovery value is roughly 4,500-5,500 MT H₂ × $1.80-2.40/kg = $8.1-13.2M/year for a 1,500 MTPD plant. The bed pays for its full installed cost ($1.4-2.0M) in roughly 2-3 months on H₂ savings alone. Annual operating cost (regeneration power, instrument air, periodic adsorbent top-up) is typically $0.4-0.6M/year, leaving an annual H₂-value net of $5-12M/year.

Parameter	Value (1,500 MTPD Haber-Bosch loop)
Purge-gas flow	10,000-14,000 Nm³/h
H₂ content in purge feed	60-72 mol%
H₂ recovery with 13X VPSA	85-92%
Recovered H₂ per ton NH₃	55-160 kg (depends on purge ratio)
Bed size	18-22 MT 13X, 4×6 mesh
Installed cost	$1.4-2.0M
Payback on H₂ savings	2-3 months
Expected bed life (well-instrumented)	5-7 years
Annual operating cost	$0.4-0.6M/year
Annual H₂-value net	$8-13M/year

The single largest variable in the TCO model is the upstream NH₃ and H₂S polishing. A plant that runs with 200+ ppmv NH₃ slip or 1+ ppmv H₂S slip will see bed life drop from 5-7 years to 18-30 months, and the adsorbent replacement cost alone ($54-84k every 2 years) wipes out 1-2% of the annual H₂ savings. The upstream polishing is the cheapest insurance in the system.

8. Featured Product: Molecular Sieve 13X from Aluminaworld

Aluminaworld Molecular Sieve 13X (Sodium Exchanged X-Type Zeolite) — 4×6 mesh (1.7-4.0 mm) and 4×8 mesh (1.7-3.4 mm) grades for Haber-Bosch loop purge-gas H₂ recovery, with bulk density 640-660 kg/m³, single-pellet crush strength ≥ 40 N, attrition ≤ 0.15 wt%, and 5,000-cycle accelerated aging report available on request. COA per ASTM D5357 sampling plan.

View Molecular Sieve 13X Product Details →

Related Aluminaworld products for the broader ammonia-synthesis gas train:

Activated Alumina for the upstream NH₃ water-wash polishing and H₂S guard
Alumina Powder for catalyst substrate applications in methanator and ammonia synthesis catalyst
Catalyst Carrier pseudo-boehmite for the iron ammonia-synthesis catalyst support
ZSM-5 Zeolite for residual NH₃ slip polishing downstream of the water-wash tower

For pseudo-boehmite grades used in the iron-ammonia-synthesis catalyst binder system, see our dedicated product page.

9. Frequently Asked Questions

Why is 13X the standard purge-gas adsorbent in Haber-Bosch loops and not 4A or 5A?

13X has a 10 Å nominal pore aperture that admits both H₂ (2.89 Å kinetic diameter) and the inerts CH₄ (3.8 Å), Ar (3.4 Å), and N₂ (3.64 Å), so all four compete for adsorption sites. The separation mechanism is kinetic, not equilibrium: at 25-40 °C and 25-35 bara, H₂ reaches a low fractional loading while CH₄ and Ar reach near-saturation, so the bed preferentially retains inerts while letting H₂ break through as the purified product. 4A (3.6 Å) and 5A (5.0 Å) exclude most inerts and would not retain them; the bed would have no separation function and H₂ would slip with the inerts.

What purge ratio keeps an ammonia loop within H₂-recovery targets?

Conventional ammonia plants operate at purge ratios of 3-6% of synthesis-gas flow to keep inert accumulation below 15 mol% in the loop. Without a purge-gas recovery bed, the H₂ lost with the purge stream is 60-180 kg H₂ per metric ton of NH₃ - the single largest hydrogen debit in the plant mass balance. A well-designed 13X purge-gas bed recovers 85-92% of this H₂, dropping the purge H₂ loss to 8-25 kg H₂ per ton NH₃ and reducing make-up H₂ demand by 3-5%.

How is the 13X bed regenerated without losing the recovered H₂?

The bed is regenerated by pressure-swing to atmospheric or sub-atmospheric pressure in a 4-bed Skarstrom cycle with feed at 25-35 bara and blowdown to 1.0-1.5 bara, or by a 2-bed VPSA cycle with vacuum to 50-100 mbar(a). The desorbed gas from regeneration is the inert-rich stream (90-95% CH₄ + Ar + N₂) which is typically sent to the fuel-gas manifold or to a once-through incinerator. H₂ stays in the high-pressure product stream during adsorption because H₂ has lower equilibrium loading on 13X than the inerts at the chosen cycle pressure.

Why is temperature swing usually added on top of PSA on ammonia purge beds?

PSA alone cannot fully desorb the heavier inerts (CH₄ in particular) from 13X at near-ambient temperature because the equilibrium loading at 1 bara and 25 °C is still 2-4 mol/kg - about 15-25% of the high-pressure loading. A small thermal-purge addition (60-80 °C regeneration gas, 3-5% of feed) drops the residual loading to < 1 mol/kg and recovers the bed's full working capacity. Modern purge beds combine PSA (for the bulk H₂ recovery) with TSA-lite (60-80 °C regeneration gas for the inerts).

What inlet H₂S or NH₃ slip damages 13X in this service?

NH₃ slip above 200 ppmv (typical of a saturated synthesis-gas cooler outlet) reacts with the extra-framework aluminum in 13X to form ammonium aluminosilicate, a non-regenerable deposit that permanently reduces working capacity by 0.5-1.5% per 1,000 hours on-stream. H₂S above 1 ppmv reacts with the sodium cations to form Na₂S, also non-regenerable. Both are mitigated by upstream polishing: a water-wash tower (drops NH₃ to < 50 ppmv) and a ZnO or activated-alumina guard (drops H₂S to < 0.1 ppmv).

Can a 4A or 5A bed replace 13X in a purge-gas duty?

No. 4A's 3.6 Å pore aperture excludes both CH₄ and Ar entirely; the purge gas would pass through the bed unseparated and H₂ would slip with the inerts as if the bed were not present. 5A's 5.0 Å aperture admits n-paraffins but not the inerts of interest in ammonia purge (CH₄, Ar, N₂ are all under 5 Å kinetic diameter; 5A is selective for n-C₆+ linear hydrocarbons). Only 13X has the right pore aperture and cation chemistry for this duty.

What is the effect of feed pressure on 13X working capacity for the purge duty?

13X working capacity rises with feed pressure but the relationship is non-linear. From 10 bara to 30 bara the working capacity for CH₄ roughly doubles (from ~2 mol/kg to ~4 mol/kg), then plateaus between 30-50 bara. Most modern Haber-Bosch loops operate the purge-gas bed at 25-35 bara (downstream of the synthesis-gas compressor discharge cooler). At < 20 bara feed, the bed is undersized; at > 50 bara feed, the marginal capacity gain does not justify the higher vessel cost.

How often should the 13X bed be replaced?

In well-designed service with upstream NH₃ and H₂S polishing, a 13X purge-gas bed lasts 5-7 years before working capacity degrades by 15% from initial. The replace-on-condition signal is rising H₂ content in the low-pressure purge stream (above 12 mol% H₂) at constant regeneration conditions. Field audits show as short as 18 months at plants with poor upstream polishing and as long as 9 years at the best-operated facilities.

Is cryogenic separation ever preferred over 13X for purge-gas H₂ recovery?

Yes - at very large plants (3,000+ MTPD NH₃) and at plants that already have a cold box in the synthesis-gas train, cryogenic separation becomes competitive. A cryogenic unit recovers 95-98% of H₂ from the purge stream but consumes 0.4-0.6 GJ per ton NH₃ of refrigeration duty. 13X is preferred at 500-2,000 MTPD plants because the capital cost is 30-50% lower and the H₂ recovery (85-92%) is close to cryogenic for typical purge compositions.

What is the role of activated alumina upstream of 13X?

Activated alumina is used as a pre-bed to remove water and heavy hydrocarbons from the purge stream before it contacts 13X. Water above 50 ppmv in the feed causes 13X hydrolysis on thermal cycling (the so-called 'dealumination' that destroys framework Si-O-Al bonds). A 200-kg activated-alumina guard upstream of the 13X main bed, regenerated in parallel, drops the feed water to < 5 ppmv and extends 13X life by 30-50%. Activated alumina is also a H₂S guard (up to 5-8% capacity for H₂S) and adds a margin of protection against sulfur slip from upstream methanator upset.

How do I size the 13X bed for a 1,500 MTPD ammonia plant?

A 1,500 MTPD ammonia plant has a synthesis loop gas flow of about 280,000 Nm³/h at 28-32 bara, with a purge stream of 8,400-14,000 Nm³/h (3-5% purge ratio). At 8 wt% 13X working capacity for the inerts at 30 bara and 35 °C, the bed needs roughly 18-22 MT of 13X for a 6-minute cycle time. The vessel is typically 2.4 m diameter × 8 m tall, with internal distributor and hold-down screens per ASME BPVC Section VIII Div 1.

Does the bed need a dust filter?

Yes. A 10-µm sintered metal or PTFE-membrane filter on the bed outlet is mandatory for protecting the downstream H₂-recovery compressor from dust carryover. 13X attrition in this service is 0.1-0.3 wt% per year under steady operation; a properly sized dust filter should be replaced annually or when ΔP across it exceeds 0.5 bara, whichever comes first. Field audits show the dust filter is the most-replaced consumable on the bed, ahead of the adsorbent itself.

10. Next Steps

If you are specifying the purge-gas bed for a Haber-Bosch ammonia loop in the 500-3,000 MTPD range, the next three actions are: (1) confirm the purge ratio and the inert composition in the loop - this sets the bed size and the regeneration scheme; (2) issue a vendor RFQ with the 14-point specification checklist in Section 6 and require written responses; (3) verify the upstream NH₃ water-wash and H₂S guard are sized and instrumented to < 50 ppmv NH₃ and < 0.1 ppmv H₂S.

For a working sample of Aluminaworld 13X molecular sieve (4×6 mesh, Na-exchanged, with full COA) for bench testing under your specific ammonia-loop purge conditions, contact us directly. We ship from Zibo, Shandong, with 15-day lead time for standard grades and 25 days for custom particle size distributions. MOQ 1 MT, free 500 g sample for qualifying ammonia-synthesis projects.

Molecular Sieve 13X — Sodium Exchanged X-Type Zeolite

4×6 mesh / 4×8 mesh grades for Haber-Bosch loop purge-gas H₂ recovery. Crush strength ≥ 40 N, attrition ≤ 0.15 wt%, 5,000-cycle report, ASTM D5357 COA.

View Product Details →

WhatsApp: +86 133 2522 2240 Email: barry@aluminaworld.com