Distinguish the product specification, a design guarantee, a sample report and the operating quality system when evaluating a recovery plant.
Four kinds of evidence answer different questions
A product specification defines what the customer requires. A process guarantee describes the performance promised within an agreed operating envelope. A laboratory report records the results for the sample submitted. An operating quality system describes how the producer controls and verifies product over time. These documents support one another, but none should be silently substituted for another.
For example, a favorable sample result can show that the submitted sample met particular reported values. It does not by itself establish every future batch, every operating condition or every unit sold by the equipment supplier. Equally, a brochure design value should not be presented as though it were a measured result from a named installation.
Define the source and the finished-product schedule
The source assessment should connect process origin with the compounds that may enter recovery. The finished-product schedule then identifies the required limits, units and methods at the agreed delivery point. A high bulk CO₂ assay does not tell the buyer whether every relevant trace constituent is controlled.
ISBT’s published resource catalog identifies bulk CO₂ guidance covering source considerations, finished-product impurities and analytical methods. The project team should obtain the applicable document and the customer’s current specification. The role of an equipment proposal is to explain how the selected source and process can be assessed against that requirement, not to treat an association’s name as a plant certificate.
Read laboratory results without changing their meaning
Keep the original unit and qualifier beside every result. A value reported as less than a stated limit is not an exact zero. A result expressed as greater than a purity value should retain that symbol. A measured number, a reporting limit and a customer maximum each mean something different.
Moisture and sulfur results deserve special care when comparing reports. Atmospheric dew point and water concentration are different reporting forms; conversion needs a suitable basis. Total sulfur and total sulfur excluding a named compound are not identical analytical scopes. The correct comparison preserves the method and reporting basis instead of placing superficially similar numbers in one unlabeled column.
Follow quality through the product path
The quality boundary may continue through storage, loading, transport and vaporization to the actual point of use. Sample locations and responsibility for release should follow that path. A customer taking bulk liquid and a brewery using vaporized gas may need different verification arrangements even when the upstream purification equipment is similar.
Online instruments can provide timely information about selected analytes, while other compounds may need laboratory verification. Calibration, sampling condition and measurement validity affect the usefulness of either result. A missing signal or a result outside the method’s capability should not become an automatic statement that the product is acceptable.
Build a purchase specification that can be tested
Ask for the feed envelope, product schedule, sampling points, test conditions and allocation of responsibilities before manufacture. The acceptance procedure should separate throughput, quality and utility performance. If the buyer needs an independent laboratory or a customer qualification step, include its timing and scope in the project plan.
Our named reference reports can support a focused conversation when their sample dates, report references and units remain visible. They should lead into the evidence required for the new project: representative source data, a defined treatment route and an agreed test plan. That gives buyers a stronger basis for confidence than a broad statement that one purity number satisfies every food and beverage application.
Common technical questions
Does an ISO company certificate certify every batch of CO₂?
No. A company-management certificate and a product-analysis result have different scopes. Review the certificate and the product acceptance evidence separately.
Can a sample report be used in a project reference?
Yes, when it identifies the sample and preserves the reported results and qualifiers. It should not be expanded into an unsupported continuous-performance claim.
Three report entries that lead to different decisions
| Hypothetical entry | What it establishes | What remains unresolved |
|---|---|---|
| Bulk CO₂ = 99.998% | The complement is 0.002%, or 20 ppm on the same fraction basis. | Which impurities make up that balance, the assay method and whether each limit is met. |
| Compound X <0.1 ppmv; customer limit 0.05 ppmv | The method reports below 0.1 ppmv. | This result alone cannot establish compliance with 0.05 ppmv. |
| Total sulfur excluding SO₂ <0.1 ppmv | A result on the stated exclusion basis. | It is not a result for total sulfur including SO₂. |
Before accepting a certificate, match sample identity, sample point, product batch, units, method, reporting limits and date to the intended delivery. Agree how measurement uncertainty is treated close to a specification limit. A report from a different source or a favorable commissioning sample cannot establish every later delivery’s quality.
Technical references
- ISBT: Bulk Carbon Dioxide quality and analytical guidance, publication BVG-00001 — The publication covers feed-gas considerations, finished-product impurity levels and analytical methods. The copyrighted publication itself was not reproduced.
- EIGA: Food and beverage gases — Industry context for food and beverage gas applications.
Apply this to your project
Open the source assessment tool to record the mass-balance and economic assumptions. For a project-specific design, share the gas analysis and product specification with us.
We publish this guide as YTC. Our editorial policy explains the source checks, AI assistance and how to report a correction. Examples do not establish a plant guarantee.