Choosing a preservative is not a matter of picking the strongest molecule on a list. It is a matching exercise: the active has to survive the formulation long enough to do its job, be acceptable in the destination market for that use category, and cost less per protected tonne than the alternatives that also survive. This article sets out the matching logic for the most common industrial actives – CMIT/MIT, BIT, OIT, bronopol, DBNPA, glutaraldehyde and the film fungicides – starting from the properties of the system rather than from the product catalogue.
1. Start from the system, not from the active
Five parameters of the formulation eliminate most of the list before efficacy is even discussed.
| System parameter | Why it eliminates candidates | What it usually leaves |
|---|---|---|
| pH of the finished product | CMIT/MIT hydrolyses rapidly above about pH 8; bronopol loses stability in alkaline conditions; DBNPA hydrolyses quickly at high pH | alkaline systems: BIT, OIT, zinc pyrithione; acidic and neutral: the full list |
| Presence of nucleophiles or reducing agents | sulfites, thiosulphates, ascorbate, thiols and amines consume isothiazolinones and bronopol | systems with reductants: BIT is more tolerant, but efficacy must be re-proven |
| Process and storage temperature | all actives degrade faster when hot; some are lost during the process step if added too early | hot processes: add after cooling, and prefer thermally robust actives |
| Required duration of protection | DBNPA is a knockdown agent that does not persist; glutaraldehyde fades; CMIT/MIT has a persistent MIT fraction | long shelf life: CMIT/MIT, BIT, or a blend |
| Use category and market | cosmetics, food-contact, toys, agricultural and marine uses have distinct restriction lists | often only two or three actives are legally available |
2. The actives, compared
| Active | CAS | Speed of kill | pH window | Sensitisation | Water solubility | Typical role |
|---|---|---|---|---|---|---|
| CMIT/MIT | 55965-84-9 | fast | acidic to neutral | high | complete | in-can workhorse |
| MIT alone | 2682-20-4 | moderate | broad | high | complete | where CMIT is restricted |
| BIT | 2634-33-5 | slow | broad, including alkaline | moderate | low, used as a solution | alkaline and long-life systems |
| OIT | 26530-20-1 | moderate | broad | high | very low | dry-film fungicide |
| Bronopol | 52-51-7 | fast against bacteria | acidic to slightly alkaline | moderate | high | bacterial control, formaldehyde-releasing |
| DBNPA | 10222-01-2 | very fast | acidic to neutral | moderate | moderate | process sanitation, not storage |
| Glutaraldehyde | 111-30-8 | fast | broad | high | complete | water treatment, metalworking |
| Zinc pyrithione | 13463-41-7 | slow | broad | moderate | insoluble | fungi and algae, dry film |
| IPBC | 55406-53-6 | slow | broad | moderate | low | film fungicide for coatings |
Two generalisations hold across the table. First, the fast-acting actives are also the ones that are consumed fastest and are the strongest sensitisers; the stable, slow actives are milder and need partners. Second, water solubility determines whether an active works in the can (aqueous phase) or in the dried film (hydrophobic phase) – a soluble active will leach out of a cured coating, an insoluble one will not protect the liquid inside it.
3. Decision logic by problem type
| Problem | First choice | Second choice | Avoid |
|---|---|---|---|
| Water-based paint, pH 7 – 9, one-year shelf life | BIT or BIT/CMIT blend | CMIT/MIT plus a stability check at the real pH | CMIT/MIT alone in a strongly alkaline paint |
| Aqueous emulsion or adhesive, pH 4 – 7 | CMIT/MIT | CMIT/MIT with bronopol for extra bacterial cover | DBNPA as the only preservative |
| Finished film attacked by mould on the wall | OIT or IPBC in the wet paint | zinc pyrithione where permitted | raising the in-can dose – it does not protect the dry film |
| Metalworking fluid, high pH, heavy load | a formulated blend, typically isothiazolinone plus a formaldehyde releaser or an amine-based actives package | glutaraldehyde where permitted | single actives in heavily loaded sumps |
| Process line fouling, biofilm on surfaces | DBNPA shock dose | glutaraldehyde programme | persistent preservatives used as cleaners |
| Cosmetic or personal care, rinse-off | restricted isothiazolinone at regulated levels | alternative actives permitted in that market | technical-grade concentrates of any kind |
| Food-contact or food-adjacent | only actives explicitly permitted for the application | – | any assumption that industrial registration transfers |
4. Blends and why they exist
Most commercial preservative products are blends, for three reasons. Spectrum: CMIT/MIT is weak against some fungi, which is why coatings use an in-can bactericide plus a film fungicide. Kinetics: a fast knockdown active combined with a slow persistent one covers the first hours after manufacture and the following year of storage. Consumption management: in a system that eats actives – amine-neutralised resins, high solids, high pH – dosing a more robust partner prevents the whole system from failing when the reactive component is used up.
The cost implication is not intuitive. Blending raises the price per kilogram of preservative but usually lowers the cost per protected tonne, because it allows a lower total dose and removes the risk of a failed batch. Conversely, buying the cheapest single active and overdosing it is often the more expensive route, both in material and in rejected production.
5. Setting the dose without guessing
The only defensible basis for a dose is a preservative efficacy test on the finished formulation, run at the real pH and solids, with the real raw materials, and read against the pharmacopoeial or industry method the customer’s market accepts. Three rules make the difference between a test that means something and one that does not.
- Test the formulation, not the active. The same 100 ppm of active can pass in water and fail in a paint, because the paint binds or consumes it.
- Age the sample before the challenge. A sample tested the day it was made has not yet lost active to the formulation. Testing after two weeks at 40 °C is closer to reality.
- Challenge with the organisms you actually meet. A preservative that passes against the standard panel can still fail against the biofilm organisms in a customer’s plant.
| Task | Rule of thumb |
|---|---|
| Convert active target to product mass | divide the active mass by the active fraction: at 14% active, 10 g of active = 71 g of product |
| Account for losses in the process | add the biocide after the hot stage and after neutralisation; assume some loss in high-solids systems |
| Verify after a stability study | re-assay the active in the aged product; a dose that cannot be measured at the end of shelf life was not a dose |
| Keep records per batch | dose, addition point, temperature, pH, and the active assay of the finished batch – these are what make a failure diagnosable |
6. Regulatory selection as a filter, not an afterthought
Restriction lists differ by use category and by market, and they eliminate candidates faster than any technical consideration. A concentrate that is fully legal as an industrial biocide may be prohibited in a cosmetic rinse-off product at any concentration; an active approved for a paint may not be approved for a food-contact adhesive; a product registered in one country may require a separate registration in the next. Three practical rules: identify the use category before shortlisting actives, obtain the registration status in writing for the destination market, and never assume that a technical-grade material inherits the approvals of a formulated branded product. In the European Union, biocidal actives are authorised under the Biocidal Products Regulation through product-type-specific approvals, and industrial use outside that framework is not permitted even if the chemistry is identical.
7. FAQ
Is CMIT/MIT being phased out?
Not as an industrial biocide. It has been progressively restricted in consumer-adjacent categories – most visibly in cosmetics, where leave-on use has been prohibited in the EU since 2017 – and some markets restrict it further. Industrial use continues under product-type authorisations, which is why the use category has to be checked before the molecule is shortlisted.
Which single preservative covers bacteria and fungi?
None covers both well at low dose. CMIT/MIT and bronopol are strong on bacteria and moderate on fungi; OIT, IPBC and zinc pyrithione are strong on fungi and weak on bacteria. Coatings are normally protected with two actives because of this.
Why did my preservative work in the lab and fail in production?
Usually because the lab sample and the production batch differ in pH, in neutralising amine, in residual reductant from the initiator, or in contamination load. The active was consumed before it could protect the product. See stability and formulation compatibility for the diagnostic sequence.
Is a higher dose always safer?
No. Above a threshold it adds cost, odour, hazard classification and regulatory burden without adding protection, and in some systems it causes phase separation or colour drift. The objective is the lowest dose that passes the efficacy test with margin.
What documentation should accompany a biocide shipment?
Batch COA with the active assays and the diagnostic impurity, SDS in the destination language, and a written statement of registration status for the intended use category and market.
Conclusion
Shortlist by pH window, nucleophile load, required protection period and use category; then prove the dose with an efficacy test on the aged formulation; then confirm registration for the market and category in writing. In that order, the choice usually resolves to a small number of options, and the cheapest of those options is genuinely the cheapest – not a false economy discovered in a rejected batch. For the chemistry and specification of the isothiazolinone option see the CMIT/MIT technical article and the CMIT/MIT-14 product page. For failure diagnosis continue with stability and formulation compatibility.