Trace Ethane Calibration Gas Blend Stability via Cylinder Passivation

Trace ethane calibration gas blends intended for analyzer linearity checks, emissions monitoring, and natural gas BTU determination require stability intervals that often exceed 12 months under transport and storage conditions. In high-pressure aluminium and 316L stainless steel cylinders conforming to DOT 3AL and ISO 11120:2015, ethane at mole fractions between 200 ppb and 100 ppm is subject to wall interactions that are not fully captured by the gravimetric preparation uncertainty defined in ISO 6142-1:2015. The cylinder wall presents a heterogeneous array of metal oxides, silanol groups, adsorbed water, and mechanical crevices; ethane, although chemically less reactive than H₂S or NO₂, has a polarizability of 4.43×10⁻²⁴ cm³ that promotes physisorption on high-energy Lewis acid centers and in micropores. At mole fractions below 1 ppm, this reversible but kinetically slow uptake can reduce gas-phase concentration by more than the combined gravimetric and analytical uncertainty budget, producing instrument calibration bias. Passivation converts the inner surface from a high-surface-energy oxide network to a low-energy, cross-linked amorphous silicon oxycarbide or fluorosilane layer, reducing active site density from roughly 5–10 sites/nm² to below 0.1 sites/nm² on vendor-certified internal finishes. Cylinder preconditioning, valve body passivation, and residual water removal are therefore not optional but are integral to ISO 17034:2016 reference material homogeneity and stability characterization. Field data from contract gas laboratories operating 10 L, 150 bar aluminium cylinders have shown that unpasivated cylinders charged with 500 ppb ethane in nitrogen can fall to 85–92% of gravimetric concentration within 72 h when residual water exceeds 5 ppmv, whereas passivated cylinders under the same filling conditions retain 99.2% ± 0.6% over 180 days at 20°C. The distinction is greatest when the cylinder is cooled, when inlet valves are opened frequently, or when the blend matrix contains carbon dioxide and water vapour at liquefied gas margins. These operational facts place the burden on the gas blend producer to qualify passivation not merely as a surface treatment but as a stability-critical manufacturing step controlled under ISO 16664:2017 storage and handling guidance.

What Surface Reactions Remove Ethane From Trace Blends During Initial Cylinder Conditioning?

The first instability interval for a freshly cleaned cylinder occurs before the ethane blend is injected, during vacuum bake-out and surface conditioning. Untreated aluminium alloy 6061-T6 and 316L stainless steel inner walls carry a native oxide layer of 2–5 nm thickness, with surface hydroxyl densities on aluminium oxide in the range of 10–15 nm⁻² after ambient humidity exposure. These hydroxyl groups and exposed metal cations function as Lewis and Brønsted acid sites that can polarize the C–H bond of ethane even at ambient temperature. Although ethane does not chemisorb to the same extent as ethylene or acetylene, differential scanning microcalorimetry and temperature-programmed desorption studies on reference alumina powders have demonstrated physisorption binding energies of 25–35 kJ/mol on high-index crystal facets, which is sufficient to delay elution and reduce headspace concentration in a closed cylinder. Capillary condensation in pits with radii below 2 nm further removes traces of ethane by irreversible occlusion if liquid water films are present. The conditioning step therefore must reduce surface moisture to a dew point below -60°C and substitute active hydroxyls with terminal methyl or fluoro groups. Passivation is not a simple rinse; it is a chemical vapour deposition or controlled liquid-phase silylation that forms Si–O–metal bonds and collapses the available surface area for ethane retention. The process is validated by measuring quasi-isosteric heat of adsorption before and after passivation; effective treatments reduce the ethane adsorption enthalpy on aluminium oxide from 28 kJ/mol to below 18 kJ/mol, a level at which wall uptake at 25°C becomes statistically indistinguishable from analytical repeatability for a 500 ppb blend over 90 days. Without this surface conversion, repeated fill–vent cycles and exposure to laboratory air during regulator changes reintroduce active hydroxyls and water clusters at the cylinder neck, causing concentration offset in the first 10–20 L of gas drawn. Such initial draw-off behaviour is a common field failure mode; the first sample after overnight equilibration can read 5–12% lower than the certified value if wall passivation and post-fill conditioning are incomplete.

Industrial passivation for hydrocarbon calibration cylinders typically begins with aqueous degreasing, mechanical or electropolishing to an inner-surface roughness Ra of 0.10–0.25 µm, followed by vacuum bake-out at 120–160°C for 4–6 h at pressures below 0.1 mbar. After cooling to process temperature, the cylinder is filled with a vapour-phase organosilane precursor such as dimethyldimethoxysilane or hexamethyldisiloxane at partial pressures between 1 mbar and 20 mbar, using a carrier gas composed of dry nitrogen or argon. The cylinder is then heated to 200–300°C for 2–8 h to drive surface condensation and crosslinking. The resulting amorphous silicon oxycarbide layer has a thickness of 100–800 nm, depending on pressure, temperature, and cycle count; two-cycle deposition reduces pinhole density and improves edge coverage at the cylinder neck. After passivation, the cylinder is purged with purified nitrogen at 80–120°C to remove unreacted precursor and reaction by-products, and final moisture is verified by cavity ring-down spectroscopy at below 500 ppb water in the gas phase. For ethane blends below 2 ppm, the cylinder valve, spindle, and burst disc assembly must also be passivated or selected from materials with low adsorption capacity, because the valve contributes 10–30% of the total wetted surface area in small cylinders. Filled stability trials are then conducted in accordance with ISO Guide 35:2017 and ISO 6143:2001 protocols, with measurements at 0, 7, 30, 90, 180 and 365 days under both ambient and 40°C transport simulation. Without these controlled parameters, variability in passivation layer thickness and surface coverage produces lot-to-lot stability differences exceeding 2% for ethane molar fractions under 1 ppm, even when the same nominal cylinder specification is used.

Passivation Layer Thickness, Roughness, and Ethane Breakthrough Humidity

The passivated layer is not homogeneous; its stability depends on thickness uniformity, residual silanol density, and the breakthrough humidity at which water vapour penetrates to the underlying oxide. X-ray photoelectron spectroscopy of passivated aluminium cylinder coupons typically shows a reduction in surface oxygen content from 55–65 at% to 15–20 at%, with carbon increasing to 25–35 at% and silicon to 10–15 at%. Atomic force microscopy confirms that the organosilane layer does not eliminate mechanical roughness but deposits conformally over electropolished features; coated surface Ra values of 0.05–0.12 µm are achievable on 6061 aluminium. The minimum water penetration threshold is evaluated by exposing passivated coupons to controlled humidity and measuring the ethane retention of a sealed test chamber; acceptable layers maintain 99% recovery at 1 ppm ethane after 14 days under 60% RH external humidity. Breakthrough humidity for single-pass organosilane layers typically lies near 45–55% RH at 25°C, whereas two-cycle and fluorinated layers extend the threshold above 75% RH and reduce ethane wall retention by a further factor of 3–5. The table below summarizes comparative stability data from a 10 L cylinder test matrix at 150 bar and 25°C for a nominal 500 ppb ethane in nitrogen blend; published data for other matrix compositions may vary with co-adsorbing species.

When Cylinder Temperature Drops Below 5°C, Ethane Adsorption Isotherms Shift Toward Irreversible Site Occupation

Temperature cycling of trace ethane blends is a severe stability test because the adsorption isotherm is nonlinear and hysteresis occurs in the presence of any residual porosity. The adsorption enthalpy of ethane on untreated or partially passivated surfaces is negative enough that lowering the cylinder temperature from 25°C to 0°C increases wall inventory by a factor of 2–5, depending on the local surface heterogeneity. When the cylinder is later warmed, the desorption rate is governed by multiple time constants; readsorption on downstream surfaces of the regulator and sample line often masks the recovery and causes an apparent concentration step. On a passivated cylinder with low site density, the same cooling cycle produces less than 0.5% deviation in headspace concentration after 24 h return to 25°C. On unpasivated aluminium, field tests have shown transient deviations of 4–10% after a 24 h excursion to -10°C, with the first sample drawn immediately after warming being the most biased. The effect is amplified when the balance gas contains water vapour, even at levels below 1 ppm; ice-like clusters on polar sites exclude ethane from the surface and alter the apparent partial pressure. Passivation therefore must be assessed under temperature cycling according to ISO 16664:2017 transport simulation, which recommends holding cylinders at -20°C and +40°C for defined intervals before returning to reference conditions and verifying the certified value by comparison with a fresh primary standard. The kinetic rate constant for desorption from unpassivated oxide sites has been reported in the range of 0.01–0.05 h⁻¹ at 25°C, meaning that full re-equilibration can take 20–100 h, far longer than the typical 30 min regulator purge interval used in field calibration. Passivation reduces the proportion of slow-desorbing sites and compresses the re-equilibration window to below 2 h, which is consistent with observed stability after cold-chain transport.

Verification of ethane blend stability in passivated cylinders requires a measurement system capable of resolving changes smaller than the combined gravimetric and analytical uncertainty. For certified values between 0.2 ppm and 10 ppm, gas chromatography with flame ionization detection and a methanizer-free hydrocarbon column provides a repeatability of 0.3–1.0% relative standard deviation when operated with split injection and molecular sieve or porous polymer columns. For lower mole fractions down to 10 ppb, a multi-bed preconcentrator with thermal desorption and cryofocusing is used to improve detection limits. Calibration curves are generated with primary reference materials prepared gravimetrically under ISO 6142-1:2015 and verified by comparison against a certified reference standard according to ISO 6143:2001. For each stability time point, at least 3 independent aliquots are drawn from the cylinder after an equilibration period of 2 h at 23°C ± 2°C, with the first 50 mL discarded to avoid sampling dead volume effects. Stability acceptance is evaluated with the criterion |x̄(t) − x̄(0)| ≤ k·√(u_t² + u_0²), where k is the coverage factor from ISO Guide 35:2017 and u represents combined standard uncertainties. A passivated cylinder filled with 1 ppm ethane in nitrogen typically yields drift values below 0.2% over 12 months, whereas an untreated cylinder may show apparent drift of 8–15% in the first 30 days, invalidating its use as a working calibration standard. Analysts must distinguish wall-induced drift from detector nonlinearity and from regulator permeation; blank tests and independent zero-gas measurements are required at each sampling interval. Analytical data from field calibration audits in petrochemical plants using 1 ppm ethane standards have shown that passivated cylinders are accepted with 95% confidence over 24 months when stored indoors, while unpassivated cylinders fail internal stability criteria within 90 days. Published data for specific blend matrices such as LNG or refinery gas is limited, but the governing physical adsorption mechanisms are transferable when co-adsorbing water and heavy hydrocarbons are controlled.

Across Refill Manifolds, Passivation Quality Determines Ethane Stability Within 30 Days

Passivation layers are subject to slow thermal oxidation and hydrolysis when cylinders are repeatedly heated during moisture dry-down or when exposed to elevated storage temperatures. The methyl groups that provide low surface energy are stable in inert gas up to approximately 300°C, but in the presence of trace oxygen and water they can be oxidized to silanol and eventually siloxane crosslinks that reduce the water contact angle from above 95° to below 60°. This degradation is evident in used cylinders that have undergone multiple 120–160°C bake-out cycles; field-reconditioned cylinders show ethane recovery drift of 1.5–3.0% over 60 days when the passivation layer thickness is below 100 nm and surface oxygen rises above 20 at%. The degradation mechanism involves the formation of silanol nests at pinholes, which re-expose underlying aluminium oxide and create local adsorption sites for ethane at the 3–7 nm pore scale. To recondition a thermally degraded passivated cylinder, the surface must be reevaluated by X-ray photoelectron spectroscopy or water contact angle measurement before refilling for sub-ppm ethane service; otherwise the cylinder may pass a simple moisture test but fail the ethane stability specification after 30 days. Production turnarounds that include multiple steam cleaning steps or hot air drying above 150°C are identified as the most common root cause of premature passivation deterioration in refillable cylinders. This is why many gas laboratories impose a maximum revalidation interval of 5 years or 20 refill cycles, whichever occurs first, for trace hydrocarbon cylinders under ISO 17034:2016 process control.

Compliance documentation for passivated trace ethane cylinders must demonstrate that the cylinder interior does not alter the certified value within the assigned stability interval. The reference material producer is required to characterize homogeneity and stability under ISO 17034:2016, which references ISO Guide 35:2017 for the statistical evaluation of stability data. Cylinder preparation and passivation are not directly addressed by a single standard; instead, the producer assembles evidence from design codes, surface treatment specifications, gas analysis method standards, and stability data. A representative compliance checklist for a 10 L, 150 bar aluminium cylinder with a 1.0 ppm ethane in nitrogen blend is compiled below. Each line item requires documented traceability to a recognized standard or validated internal procedure.

One operational boundary consistently observed on production filling lines is the incompatibility of passivated ethane blends with moisture-introducing sample systems and with certain elastomer seals. Even if the cylinder interior is passivated, a regulator with nitrile or neoprene elastomers can introduce moisture, plasticizer, or hydrocarbon impurities that co-adsorb and distort ethane concentration during draw-off. Production-scale audits at a natural gas custody transfer facility using 200 ppb ethane calibration standards found that changing from an unpassivated cylinder to a passivated cylinder reduced calibration offset by 5.2%, but only when the sample line was simultaneously replaced with electropolished stainless steel tubing and the regulator was purged for 20 min at 200 mL/min. The same audit showed that passivated cylinders stored with their valves open to a common manifold recovered their certified value within 4 h, whereas unpassivated cylinders required 48–72 h recovery after the same exposure. Incompatibility boundaries include prolonged storage above 40°C, external humidity above 85% RH without cylinder cap protection, and use of chlorine-containing cleaning agents near the valve inlet, which can degrade methyl-terminated silane layers. Operational specifications for the fill manifold must also include a dry-down sequence after cylinder changeover to hold residual water below 0.5 ppm and an oxygen content below 1 ppm before introducing trace ethane. For blends containing ethane and carbon dioxide at concentrations near the saturation pressure, passivation is necessary but not sufficient; phase condensation at low temperature must be separately controlled by blend design. Published data for specific passivation suppliers and for ethane blends in biomethane or high-BTU natural gas matrices is limited, so stability qualification must be performed on the final blend composition rather than inferred from nitrogen-matrix studies. The technical judgement required is therefore to treat passivation as part of the analytical instrument system, not as a one-time cylinder treatment, and to revalidate ethane stability after any cylinder repair, valve replacement, or exposure to uncontrolled atmosphere.