Can TCD and GC provide accurate concentrations for many types of mixed gases?

[CAFontaine]

I am trying to purchase accredited primary cal gas but i don't know how to interpret the Scope of calibration provided by the gas provider. I am attaching a picture of the scope (specifically interested in the gas concentration portion and GC & TCD) and a picture of the gas mixtures in question. Thank you so much, it is hard to find accredited calibration gases for ISO 17025:2017 accredited calibrations. Also if my previous question is too broad, if I ordered 8% CO with the balance of Nitrogen would the attached COA under the attached scope of accreditation provide me traceability to NIST?

[CAFontaine]

I think I have this run to ground but it is causing serious concern. For a mixture to be chemically traceable to NIST the gas must be compared with a NIST SRM (Standard Reference Material) for that mixture. If gravimetric is used the traceability is in Mass only and not Chemically traceable. So for a Calibration Lab to perform an accredited calibration on a gas meter it would have to use a standard compared against the NIST SRM. My concern is that Labs, even for accredited calibrations, are already not using primary standard test gases compared to NIST SRM's, they are hopefully at least using gravimetrically produced primary standard gases accompanied by a COA and covered by a Certificate of Analysis, and they are possibly using run of the mill standard gases (not good). Then you have to wonder what test gas is being used when an accredited calibration is not requested. I know currently the US Navy does not require accredited calibrations on gas meters.

[scooke · Accepted solution]

Dear Chris,

Specialty gasses were a major part of my career. The example certificate is excellent, provides traceability to NIST standards, and can be used as a “Primary Standard” in gas analyses. Gravimetric preparation is the most accurate and reliable method for this type of mixture. The TCD is more than sufficient for confirming the correct blend (against another Primary Standard).

A good source of information and recommendations for accuracy, precision, relative error and traceability is in the EPA documents for emission gas analysis requirements.

https://www.epa.gov/emc

See also this excellent overview: https://thetruthaboutforensicscience.com/nist-traceable/

NIST also published a free booklet on analytical critical properties. The same information is available in their online manual here: https://www.itl.nist.gov/div898/handbook/mpc/section3/mpc3.htm

The most important thing for good measurements is knowing what your “error” is. You can’t do better than the comparative standard that you are using. The whole traceability issue is all about verifying the error propogation to ensure the most valid results and error ranges.

Best regards,

Steven

[scooke]

Dear Chris,

I am not really sure what your concern is here.  NIST is non-regulatory, so any "proof" of a measurement is only one that is demanded by a customer.  "Chemicals" are NOT traceable to NIST - who says so? (Not NIST).  They may be assayed by various methods.  The METRICS are maintained by NIST as part of the international system for metrics and calibrations.  I CAN assure you that only the highest quality pure gases are used in the preparation of certified primary standards - as well as all secondary standards.

Again, the important consideration is just knowing what the limits (error) of a measurement is.  You cannot simply improve that by proclamation.  To create traceability you MUST use FUNDAMENTAL constants!  That is why we use mass instead of "chemical composition".  You can calculate the error introduced to a mixture calculation from the amounts of impurities that may be present.  For example, WHAT is the actual delivered difference in a mixture of 8% CO in N2 when the CO is 99.99% pure or 99.999% pure?

Gas meter calibration is a separate application and involves its own series of primary to secondary certifications.  Back to reality, if there is no requirement for accredited calibration there is already a pretty good reason for that. IF that reason changes, then the requirements will change as well.  If my acceptable "error" range of a nominal measurement is broad enough I can essentially use any available gas to "calibrate" my meter.  In those cases, purchasing a Gold-Star Primary Standard with full NIST traceability is probably a big waste of money!

Always be aware of who is setting the requirements and who is providing the verification.  They are not usually the same parties.  In fact, where a value IS of particular importance, a formal contract is usually made that defines the requirements and acceptable test methods and standards.  If one doesn't exist, it probably isn't really that important.

Best regards,

Steven

[CAFontaine]

I am concerned when the requirement is for an accredited calibration that some Labs are still using $34 Calibration Gases, when, because it is accredited calibration, the standard must be traceable to SI units. You have provided very good information and I appreciate it. My statements regarding Gas not being traceable to NIST, was just the fact that traceability is through Mass. To establish chemical traceability to NIST you would have to compare the gas to the NIST Standard Reference Material (SRM) for that gas. Typically a 6L Gas SRM costs about $3600. The gravimetric traceability does meet our Accrediting Bodies requirements for traceability of the calibration gas.

All your other points are well taken, understood, and as I mentioned extremely helpful. You speak from a position of having seen calibration gases produced and are proud of the industry. I have never been involved in that and view from an Accredited Calibration Lab trying to provide the most accurate calibrations I can that meet my customers needs. I think that is a good thing and you have made me feel better about the gas producing industry. In my 2 month struggle to find a calibration gas that met my accrediting bodies requirements, I could only get distributors to talk with me and not producers. Thanks to you and Dr. Hodges at NIST, this country boy is a lot smarter about the measurement area now.

Thanks again Steven

[scooke]

Dear Chris,

Thanks!  I was also an auditor for that sort of thing for much of my career (still certified, but not active these days).  I can assure you that obtaining and retaining an Accreditation as an Analytical Lab is not a trivial matter!  CLAIMS to such - or disreputable "accreditors" - may be out there, but that is also on the due diligence of the client. The American Association for Laboratory Accreditation is a good place to start. American Association for Laboratory Accreditation

IF a vendor was using sub-standard standards (and 'price' is not really an indicator) that should be reported to whoever issues their certification.  If it is an assumption or unwarranted concern, be careful how you state it.  Also, be aware that it is rarely required to use a direct NIST SRM!  Those are used by the vendors to calibrate their secondary standards.  Again, carefully consider WHAT accuracy you need, and WHY!  MOST of the gas calibration business is based on the secondary standards (those derived by comparison to the NIST SRMs).  In my experience, more error in final values come from subsequent mathematical application or poor lab technique by the user than will ever be "traceable" to the standards used. Oftentimes, the standard accuracy is completely excessive for the reality of a particular experiment or analysis.  The REAL difficulty in most analyses is accurate and representative SAMPLING, not the "quality" of the analytical standard.

I've always enjoyed chemical analysis because it is so close to other philosophical questions of life.   "How do we know what we know?"

Best regards,

Steven