A calibration gas is only as good as the number printed on the label — and that number has to mean something the day a regulator, a customer, or your own quality manager asks you to prove it. The difference between a blend you can defend and a blend you merely bought is mostly in the specification you sent.
This is the spec process we use for our own standard gas line, and the one we send to customers who ask us how to brief a supplier — it works identically for any producer, ours included.
target gas and exact concentration, balance gas and its grade, the expanded uncertainty you’ll accept per component, the traceability chain, the cylinder, the shelf life, and the CoA format. Seven fields. A supplier who quotes a price before all seven are on the table is selling you a cylinder, not a reference.
What the gas is actually doing for you
A calibration gas — standard gas, reference mixture, depending on where you’re standing — is what tells your instrument what “correct” looks like. Portable gas detectors, fixed combustible and toxic gas systems, GC and FID and NDIR analyzers, emissions monitoring stations: all of them inherit the accuracy of their last reference. That’s why two properties dominate the purchase, and everything else is subordinate to them:
The concentration has to be right — which is decided by the blending method and certified on the CoA. And it has to stay right — from blend date, through transport, to the last day of shelf life.
A quick map of who buys what:
| Use case | Typical targets | Typical range | Balance gas |
|---|---|---|---|
| Occupational safety, portable detectors | CH₄ (LEL), O₂, CO, H₂S | O₂ in % vol; toxics in ppm | Synthetic air (detectors) / N₂ |
| Environmental emissions | SO₂, NOx, CO, CO₂ | ppb–ppm | Synthetic air or N₂ |
| GC / process analyzer calibration | H₂, CO, CO₂, C₁–C₄ | ppm | N₂ or He, matched to the instrument |
| Fuel / natural gas analysis | CH₄ + C₂H₆ + C₃H₈ + N₂ + CO₂ | % vol, multi-component | custom |
| Leak detection, specialty | per application | ppb–ppm | per application |
The seven fields, and what goes wrong when each is missing
1. Target gas(es) and exact concentration(s)
“1000 ppm CO in synthetic air” is a specification. “CO calibration gas, standard” is an invitation to a dispute. Multi-target blends — 1000 ppm CO plus 500 ppm H₂S plus an O₂ deficit, the kind a multi-gas detector fleet actually runs — list each component separately, each with its own uncertainty requirement. And on the zero side: zero points use zero air or N₂, and the span point should sit near the instrument’s alarm or working range, not under its detection limit where it proves nothing.
2. Balance gas — and its grade
Synthetic air (roughly N₂/O₂/Ar/CO₂) when the instrument’s oxygen reading is part of the measurement, because calibrating an O₂-sensitive detector on pure nitrogen teaches it the wrong zero. Nitrogen as the default for GC work and most toxic spans. Helium where the detector or the FID head demands it. And the balance gas itself has a grade — 99.999%+ N₂ for electronic work — because the balance’s impurities have to sit far below your targets or the targets are fiction.
3. Expanded uncertainty, per component
The CoA must state the expanded uncertainty at k = 2 for each target. A useful benchmark: for a 100 ppm blend, a competent lab delivers uncertainty on the order of a few percent of the target value. As you push into ppb, the uncertainty budget tightens and the price follows — so state the uncertainty you’ll accept, and let the supplier tell you which blending method it costs.
4. Traceability
Who does the supplier’s number trace back to? The cleanest answer is an ISO/IEC 17025-accredited production scope covering gas reference materials, with reference to a national metrology institute (NIST in the US, NIM in China). Where a producer is ISO 9001-only, ask which independent lab verifies the CoA — and get the lab’s name in writing. ISO 6142 is the method framework for static dilution reference gases, and it’s a phrase that belongs in your specification.
5. Cylinder, pressure, valve
The standard sizes run 4 L, 10 L, 15 L, 20 L, 40 L, 47 L, 50 L. A detection-equipment factory doing weekly spans across a fleet usually lands on 20 L as the economics; a single analyzer is happier with 4–10 L. Match the valve to your regulator — CGA in the US, DIN or ISO elsewhere — the same rule as any imported gas, detailed in our import checklist.
6. Shelf life, in writing
Static blends typically hold 12–24 months from blend date, but oxygen-containing mixes, reactive targets and ppb components get shorter guaranteed windows. The shelf life belongs on the label and the CoA, with storage orientation and temperature guidance, and with a replacement clause for out-of-spec cylinders.
7. The CoA format, agreed before the first shipment
Blend date, cylinder serial, each target with concentration and expanded uncertainty, balance gas, reference conditions — 20 °C/101.325 kPa or 25 °C, stated explicitly, because the two bases give different numbers for the same physical gas — and the verification method. For importers the CoA travels with the dangerous goods paperwork, so its quality affects two processes at once.
How the blends get made — this is where price is set
| Method | The idea | Best for | Realistic accuracy |
|---|---|---|---|
| Static (multi-pressure reduction) | Cylinders diluted in sequence across pressure stages | ppm range, most detector calibration gas | roughly 1–3% relative |
| Gravimetric | Weighing exact masses of each component | ppb–ppm, high-value reference standards | under 1% relative |
| Dynamic (flow blending) | Continuous mixing through mass flow controllers | high volume, or ppb work on demand | roughly 1–2% relative, stable output |
Ask the supplier which method they used on your blend, and why. The answer also tells you their price: gravimetric ppb work costs what it costs because it’s slow, and a quote that prices it like a static ppm blend is telling you something about where the number on the CoA came from. If the method can’t be named, the uncertainty claim is unverifiable, and this whole guide is academic.
Blend in-house or buy premixed — the honest version
This is the question a detection-equipment factory asks us most, and the answer has a number in it:
| Factor | In-house blending | Buying premixed |
|---|---|---|
| Upfront | High — blending rig, mass flow controllers, gravimetric balance, QC lab | None, per-cylinder price |
| Unit cost at volume | lower | higher |
| Time to first cylinder | months — rig plus method validation | days to weeks |
| QC burden | you own the 17025-style chain | the supplier’s accreditation carries it |
| Custom range | unlimited | the supplier’s capability |
| Where it wins | roughly 500+ cylinders/year, or exotic ppb reference standards | everything else |
For most detection factories the pattern that works: premixed for the detector span gases — fast, accredited, per-lot CoA — and in-house blending reserved for the high-volume or unusual concentrations that actually dominate the cost curve. The two coexist fine; the mistake is building the rig before the volume justifies it.
Five checks before you choose a calibration gas supplier
- Method disclosure — can they name the blending method and the uncertainty budget for your exact specification?
- Accreditation scope — ISO/IEC 17025 preferred, or ISO 9001 plus a named independent verification lab; see the certificate and its scope, not the wall frame.
- Reference conditions stated — the CoA in 20 °C or 25 °C basis, explicitly.
- Shelf life in writing — label and CoA, with the replacement clause.
- Export package, if importing — GHS SDS, UN classification, per-lot CoA, ISPM-15 marks. Full detail in the import guide.
A producer with a dedicated standard gas line is the natural candidate to run through these checks. Ours has been in production since 2018, covering detection, environmental and process calibration blends under ISO 9001 with per-lot CoAs — but run the five checks on us the same way you’d run them on anyone else. That’s the point of the list.
FAQ
Where do I source certified calibration gas for gas detection instruments?
From a producer or lab that states the blending method, the expanded uncertainty per component at k = 2, and traceability to a national metrology standard — ideally with ISO/IEC 17025 in scope. International majors and specialist producers (including Chinese producers with dedicated standard gas lines, like ours) all qualify if they pass the five checks above. The certificate is the shortlist; the method disclosure is the decision.
What concentration ranges do calibration gases come in?
Sub-ppb (gravimetric work) through ppm (where detector calibration lives) up to percent-level multi-component mixes. For portable detectors: O₂ in percent volume, combustibles as %LEL or ppm CH₄, toxics — CO, H₂S, SO₂, NO₂ — in ppm, with zero points in zero air or nitrogen.
How do I specify a custom ppm or ppb blend?
Seven fields: targets with exact concentrations, balance gas and its grade, expanded uncertainty per component, traceability requirement, cylinder and valve, guaranteed shelf life, CoA format. Send it as a one-page spec. A serious supplier answers with the blending method and the uncertainty budget, not just a price.
Is 100 ppm CO in air the same as 100 ppm CO in nitrogen?
The CO is the same. The instrument’s answer can differ, because oxygen presence affects some sensors and detectors. Match the balance gas to the instrument’s calibration procedure — synthetic air for oxygen-sensitive detectors, nitrogen for most GC work. Getting this one field wrong silently recalibrates your whole fleet to the wrong zero.
How long does a calibration gas last?
12–24 months from blend date for static ppm blends is typical; ppb components and reactive targets get shorter guarantees. The number has to be on the label and the CoA, and a good supplier replaces out-of-spec cylinders under the warranty clause — which is why the clause belongs in the PO, not in a conversation.
Calibration gas vs. standard gas vs. reference gas — same product?
Mostly it’s regional and usage vocabulary: “standard gas” and “reference gas” (ISO 6142′s language) emphasize metrological traceability; “calibration gas” emphasizes the use. The purchase requirements — method, uncertainty, traceability, CoA — are identical. Quote the spec, and the terminology stops mattering.
Can I get a multi-component blend — CO plus H₂S plus an O₂ deficit — in one cylinder?
Yes, and for multi-gas detector fleets it’s the normal configuration. State each component with its concentration and uncertainty in the spec. Dedicated blending lines make multi-component mid-volume work economical, which is exactly the gap between a gas trader and a standard gas producer.
Where this leaves you
Write all seven fields, demand the method and the uncertainty budget in writing, and buy premixed until your volume genuinely justifies a rig. Done that way, a custom calibration gas stops being a can with a label and becomes a metrological reference you can defend in any audit — which is, finally, the whole reason it exists.
Sources
- ISO — ISO 6142 (reference gas preparation) and ISO/IEC 17025: iso.org
- NIST — reference materials and traceability: nist.gov
- OSHA — gas detection and instrument calibration requirements: osha.gov
- UNECE — UN Model Regulations, transport classification: unece.org
- Manufacturer public websites (calibration gas capabilities): newradargas.com and the producers in our Chinese specialty gas comparison
Post time: Oct-27-2026