Hydrogen Compression Technology Selection

Hydrogen is the hardest common gas to compress. Its very low molecular weight (2 g/mol), high solubility in steel, and tendency to leak through anything make technology selection the most consequential decision in a green-hydrogen or refining-expansion project. This guide lays out the four machine families, where each fits, and the traps that turn a tidy capex estimate into a multi-year commissioning disaster.

Selection matrix -- pick before you size

Decision guide

Flow < 500 kg/day and P > 200 bar (filling stations, industrial gas)

Diaphragm compressors win. Oil-free, gas-tight, capable of 99.9999% purity with no lubricant carryover. The penalty is diaphragm replacement every 1,000-2,000 hours -- budget it and have spares on site. Single-stage pressure ratio up to 10 makes trains short. Avoid if you have to compress wet hydrogen (pre-dry it; condensate destroys diaphragms).

Flow 500-50,000 kg/day at any pressure

Lubricated reciprocating compressor is the workhorse. It tolerates dirty inlet gas, starts against load, and the valve/ring technology is mature from 50+ years of refinery makeup-hydrogen service. If the customer cannot tolerate lubricant carryover -- PEM electrolyzer stacks, fuel cells, semiconductor process -- specify non-lubricated with PTFE rider rings and longer cooling jackets, and budget for 2x the ring inspections. Do not use oil-flooded screw compressors; hydrogen dissolves the oil and blinds the separator.

Flow > 50,000 kg/day (refinery hydrogen makeup, ammonia synthesis, large SMR)

Centrifugal, multistage, usually geared or integrally geared. With H2's MW of 2, tip speeds have to hit 350-450 m/s per stage to develop meaningful head -- you need ceramic-coated impellers, rotor dynamics checked per API 617 Annex I, and active magnetic bearings on anything above 15 MW. Expect 10-14 stages per casing to make 70 bar from 20 bar suction. Centrifugals never start against full downstream pressure -- you need a blowdown sequence and a large recycle cooler.

Edge cases

• Ionic-liquid compressors (Linde patent) replace the piston-cylinder gap with a liquid; very high pressure ratios with near-isothermal compression. Pay a premium, but zero leakage and no embrittlement-prone reciprocating parts.
• Electrochemical hydrogen compressors (EHC) are promising for small-scale pure-H2 applications but have limited vendor ecosystem as of 2026. Watch the space.

The embrittlement problem -- material selection

Rule: in H2 service, no carbon steel above 350 bar, and no steel above 22 HRC hardness anywhere. ASME B31.12 PL-2.1.4 is non-negotiable; weld hardness traverse tests are mandatory.

Typical materials by part:

• Cylinders and heads: ASTM A350 LF2 for low P; A322 4140 for high P with strict hardness control; forged 316L stainless for > 500 bar.
• Piston rods: 17-4PH H1150 or nitrided 4140; avoid chrome plate (micro-cracks trap H).
• Valves: PEEK or hardened 420 SS; TI/Nb superalloys for high-speed reciprocators.
• Seals: PTFE-carbon composite or filled PEEK; buna-N embrittles fast in H2 and leaks within days -- never use it.
• Piping: stainless 316L preferred; carbon steel SA-106 Gr B acceptable up to 350 bar with B31.12 Hf de-rating.

Pulsation and vibration

Reciprocating H2 compressors are pulsation nightmares. The low molecular weight means high acoustic velocity (> 1,300 m/s), so acoustic natural frequencies of piping fall near compressor multiples. Always do a pulsation study per API 618 7th Ed. Design Approach 3 (full acoustic + mechanical analysis) before purchase. Expect dampers, orifice plates, and branch-connection bracing -- and budget for them in the estimate.

Aftercooling, drying, and downstream

• Always aftercool between stages to ≤ 40 C before the next stage -- otherwise the next suction is too hot, polytropic head skyrockets, and you're heating up to 350+ C in the final stage.
• Drain condensate every stage -- wet H2 accelerates embrittlement.
• Pulsation bottles/dampers sized per API 618 Approach 3 -- do not accept vendor "standard" bottles.
• Downstream KO drum + coalescer + molecular-sieve dryer if feeding fuel cell or PEM stack.

Quick checklist before you sign a PO

1. Flow, P1, P2, T1 specified at minimum, rated, and normal -- and max future case?
2. Is the gas pure H2, H2/CH4 mix, or H2/CO2? MW and k can shift enough to move the surge line.
3. API 618 Approach 3 pulsation study committed (line item in vendor scope)?
4. Hardness ≤ 22 HRC everywhere, documented per B31.12 and NACE MR0175?
5. Operating envelope covers startup, normal, turndown, trip -- and recycle path sized for trip case?
6. Leak-test plan -- full helium-leak test at 1.1x MAWP, acceptance 1e-6 scc/s?
7. If centrifugal: anti-surge control committed (see companion KB article), and recycle valve ≤ 2 s stroke time?

References

• API 618, 7th Ed. -- Reciprocating Compressors for Petroleum, Chemical, and Gas Industry Services.
• API 617, 9th Ed. -- Axial and Centrifugal Compressors.
• ASME B31.12, 2019 -- Hydrogen Piping and Pipelines.
• NACE MR0175 / ISO 15156 -- Materials for Use in H2S-Containing Environments (also referenced for H2 service hardness limits).
• DOE Hydrogen Program Record 16003 (2020) -- Compressor performance data for LDV fueling stations.
• Di Bella, F. "Development of a Centrifugal Hydrogen Pipeline Gas Compressor" -- DOE FE0031550, 2020.
• Linde Ionic Compressor technical brochure, 2019.