Microbial Limits for Botanicals & Dried Herbs: Buyer Guide

Key takeaways

• Microbial limits depend on category and end use, not one universal number. A dried herb sold as a boil-water infusion, a solvent extract, and a food seasoning are judged against different acceptance criteria, so the RFQ must name the framework before a lab result means anything.
• European Pharmacopoeia 5.1.8 sets three herbal categories with very different ceilings. Category A (herbal teas prepared with boiling water) tolerates TAMC up to 10⁷ CFU/g, while extract-grade Category B caps TAMC at 10⁴ CFU/g and adds a bile-tolerant Gram-negative limit.
• Pathogen criteria are absolute, not gradual. Across food and pharmacopoeial frameworks, Salmonella must be absent in 25 g, and E. coli and bile-tolerant Gram-negatives are controlled tightly even where total counts are allowed to run high.
• Untreated botanicals naturally carry high counts, so decontamination is a trade-off. Steam, gamma irradiation and ethylene oxide differ sharply on residues, EU legality, labelling and aroma damage; ethylene oxide is a banned fumigant in the EU.
• Arovela should be judged on documented lot control, not invented claims. The relevant Arovela systems are ISO 22000, ISO 9001 and ISO 27001; HACCP, GMP, organic or retailer schemes are buyer-side requirements unless separately evidenced.

Introduction

Microbial limits are where a beautiful-looking botanical lot quietly fails an audit. Dried leaves, flowers, roots and seeds are grown in soil, dried in open air, handled by many pairs of hands and shipped across borders — every step adds bioburden. A chamomile flower or a sage leaf can look perfect and still carry a total aerobic count in the millions, plus yeasts, moulds and the occasional pathogen. The buyer's job is not to demand a sterile product that does not exist in nature; it is to define which counts are acceptable for the intended use and to prove it lot by lot.

This guide is written for procurement, QA and regulatory teams importing Turkish dried herbs, botanicals and dried fruit into the EU and Ukraine. It explains the two frameworks buyers collide with most often — European Pharmacopoeia 5.1.4 and 5.1.8 for herbal-medicinal thinking, and food-hygiene expectations such as the European Spice Association minima for food use — and it shows how to turn those into COA and RFQ language that prevents disputes. For adjacent contaminant controls, read the Arovela guides on heavy metals in botanicals and dried fruit, aflatoxin and mycotoxin limits and reading a botanical COA.

Why untreated botanicals carry high counts

A dried botanical is not a manufactured chemical; it is an agricultural surface. Leaves and flowers have enormous surface area, they collect soil splash and airborne dust, and they are frequently harvested and dried without a validated kill step. Roots and rhizomes start life below ground, in direct contact with the richest microbial reservoir on a farm. Even careful open-air or shade drying only lowers water activity — it does not eliminate the organisms already present.

The result is that raw, untreated dried herbs commonly show a total aerobic microbial count (TAMC) well above what a naive buyer expects. Total counts in the range of 10⁵ to 10⁷ CFU/g are routine for untreated leaf and flower material, and yeasts and moulds are almost always present because mould was detectable in the large majority of survey samples in the published literature. This is normal microbiology, not automatic evidence of a dirty supplier. What separates a controlled supplier from a careless one is not a magically low raw count; it is drying discipline, water-activity control, storage hygiene, and an honest declaration of whether the material was decontaminated before shipment.

Two consequences follow. First, a specification that demands pharmaceutical-grade low counts on untreated herbal drug material is often unrealistic — the pharmacopoeia itself acknowledges this. Second, because raw counts are naturally high, the pathogen tests (Salmonella, E. coli, bile-tolerant Gram-negatives) carry most of the safety weight. A lot can legitimately show a high TAMC and still be safe if the pathogen and indicator criteria pass.

What the numbers actually measure

Before comparing limits, buyers should be clear on what each test reports, because a COA that lists a limit without a method is not verifiable.

• TAMC — Total Aerobic Microbial Count. The total viable aerobic bacteria per gram, enumerated per Ph. Eur. 2.6.12. It is a hygiene and freshness indicator, not a direct safety verdict.
• TYMC — Total combined Yeasts and Moulds Count. Yeasts and moulds per gram, also per 2.6.12. High TYMC is a spoilage and, indirectly, a mycotoxin-risk signal — mould counts and aflatoxin risk travel together in susceptible matrices.
• Bile-tolerant Gram-negative bacteria. An enumeration (Ph. Eur. 2.6.31) used as an indicator of enteric-type contamination and poor hygiene. It is a graded limit, not a pure absence test.
• Escherichia coli. A faecal-contamination indicator; depending on framework it is an absence test (in 1 g) or a low numeric ceiling.
• Salmonella. A pathogen; always an absence test over a defined sample mass (commonly 25 g). There is no acceptable non-zero level.

Enumeration criteria are interpreted with a tolerance factor: an acceptance criterion of 10¹ CFU allows a maximum count of 20, 10² allows 200, and 10³ allows 2000. So "10² CFU/g" on a COA is a band with a defined upper bound, not a single point.

European Pharmacopoeia 5.1.8: the three herbal categories

The framework most buyers of medicinal-grade botanicals must understand is European Pharmacopoeia general chapter 5.1.8, "Microbiological quality of herbal medicinal products for oral use and extracts used in their preparation." It groups finished herbal products into three categories, A, B and C, based on how they are prepared and processed, and sets recommended acceptance criteria for each. The distinction is decisive: the same physical herb can sit in Category A or Category B depending on whether it is brewed with boiling water or extracted.

• Category A covers herbal drugs intended for infusions and decoctions prepared with boiling water — classic herbal teas. Because the consumer effectively pasteurises the material in the cup, the ceilings are the most permissive.
• Category B covers products such as extracts and herbal drugs where the method of processing (for example extraction) reduces organisms below the Category-B limits. Extracts default to Category B.
• Category C covers products where processing — for example extraction with low-strength ethanol or non-boiling water, or low-temperature concentration — would not reduce organisms sufficiently to meet Category B.

The gap between the columns is the whole point. Category A tolerates a TAMC of 10⁷ CFU/g precisely because a boiling-water infusion is expected; the maximum acceptable count is 50,000,000 CFU/g. Category B, aimed at extracts and processed material, tightens TAMC to 10⁴ CFU/g and introduces a bile-tolerant Gram-negative ceiling of 10² CFU/g and an E. coli absence-in-1-g requirement. Notably, Salmonella must be absent in 25 g in every category — the pathogen line never moves.

The chapter also states, importantly for buyers, that for some herbal materials the TAMC, TYMC and bile-tolerant Gram-negative criteria cannot be met because of the typical level of contamination, and that higher acceptance criteria may then be applied on the basis of a documented risk assessment. This is the pharmacopoeia's own acknowledgement that untreated botanicals run high — and it is exactly the clause a supplier should reference rather than quietly failing a spec.

Where 5.1.4 fits

Chapter 5.1.4 covers non-sterile pharmaceutical preparations more broadly and is where you land if a botanical is used as a substance for pharmaceutical use rather than as a food-facing herbal product. For substances for pharmaceutical use it sets TAMC 10³ CFU/g and TYMC 10² CFU/g, with specified-organism absence requirements depending on the route of administration. Buyers of extract material destined for pharmaceutical dosage forms should confirm which chapter their finished-product regulatory route invokes, because 5.1.4 is materially stricter than Category A of 5.1.8.

EU food use versus herbal-medicinal expectations

A large share of Turkish botanicals — culinary herbs, tea ingredients, spice-adjacent leaf and seed material — are placed on the EU market as food, not as herbal medicinal products. Food use is not governed by 5.1.8. Instead, buyers typically apply food-hygiene reasoning and industry minima, most commonly the European Spice Association (ESA) "Quality Minima" expectations for dried herbs and spices, together with the general EU food-safety obligation that food must not be unsafe.

The practical food-use targets buyers reference look different from the pharmacopoeia:

The headline is that food-use TAMC is usually a commercially agreed number rather than a single legal ceiling, while the pathogen and indicator tests (Salmonella absent in 25 g, E. coli and yeast/mould ceilings) do the safety work. A finished-food brand — especially one making ready-to-eat or child-facing products — may set tighter limits than the ESA minima, and that must be stated in the RFQ before sampling. A supplier who answers "EU compliant" without naming food-use versus herbal-medicinal framing, the applicable limits and the intended matrix has not answered the question. For food-safety-management context, Arovela operates under ISO 22000, which supports hygiene management but is not itself a microbial certificate for any given lot.

Decontamination: steam vs irradiation vs ethylene oxide

Because raw counts run high, many botanical lots are decontaminated before shipment. The three dominant options are not interchangeable — they differ on residues, EU legality, mandatory labelling, aroma and colour impact, and cost.

Steam treatment

Saturated-steam treatment is the most common decontamination method used for herbs and spices in the EU. Its commercial advantages are that it leaves no chemical residue, requires no special consumer label declaration, and is broadly accepted by retailers and consumers. The trade-off is physical: steam adds heat and moisture, so it can flatten volatile aroma, dull colour and — if the material is not re-dried and re-stabilised properly — raise water activity and re-spoilage risk. For aromatic material such as chamomile, oregano leaf or sage, the buyer should compare treated and untreated samples in the final application, because a microbiologically cleaner lot with a weaker aroma may be the wrong trade for a premium tea.

Gamma irradiation

Ionising-radiation treatment is legal in the EU but tightly constrained. Under Directives 1999/2/EC and 1999/3/EC, the only category on the EU-wide positive list is "dried aromatic herbs, spices and vegetable seasonings," with a maximum permitted dose of 10 kGy. Any irradiated food or ingredient must be labelled "irradiated" or "treated with ionising radiation." Irradiation is highly effective against microbial load and leaves no chemical residue, but the mandatory labelling and low European consumer acceptance make it commercially unattractive for many retail-facing programmes. It is also unsuitable for materials outside the positive-list category, and buyers should confirm both the target matrix and the destination-market stance before requesting it.

Ethylene oxide (ETO)

Ethylene oxide is a banned fumigant in the EU — it has not been permitted as a pesticide/fumigant on food since 1991, being classified as a mutagen, carcinogen and reproductive toxicant. Under Regulation (EC) No 396/2005 (as amended), the maximum residue level in spices is 0.1 mg/kg, expressed as the sum of ethylene oxide and 2-chloroethanol, calculated as ethylene oxide. The 2020–2021 RASFF ethylene-oxide incident, which began with sesame and spread across many ingredient categories, showed how a single fumigated input can contaminate an entire supply chain — and 2-chloroethanol can persist as a degradation product or via cross-contact even without direct fumigation. For EU-bound botanicals the practical rule is simple: ETO-fumigated material is not acceptable, and the RFQ should say so explicitly and require an ETO/2-CE residue statement where risk exists.

Sampling: why the plan matters as much as the limit

A microbial result is only as good as the sample behind it. Botanical contamination is heterogeneous — organisms cluster on some flowers, some corners of a bale, some sub-lots — so a single grab sample can badly under- or over-represent a lot. Buyers should require a documented sampling plan: how many increments were taken, from how many bags or bales, how they were combined into a composite, the sample mass tested (critical for a 25 g Salmonella determination), and aseptic handling to avoid post-harvest cross-contamination during sampling itself. For pathogens, the sample mass is part of the specification: "Salmonella absent" is meaningless without stating it was tested in 25 g. First orders, new origins and new crop years justify tighter sampling; a stable supplier history can support risk-based frequency later, but the pathogen tests normally stay in every release. The same lot discipline that governs pesticide-residue management applies here — one composite, one lot code, one release file.

COA and RFQ language that prevents disputes

A weak COA reports "microbiology: pass." A defensible COA reports each parameter with its method, numeric result, the applicable limit and the sample mass. Buyers should reject documents that cannot be tied to the carton label, invoice and packing list.

Example RFQ wording buyers can adapt:

"Supplier shall provide, per lot, TAMC and TYMC (Ph. Eur. 2.6.12), bile-tolerant Gram-negative bacteria and E. coli (2.6.31), and Salmonella (absence in 25 g). Results shall state method, numeric value, acceptance limit and the microbiological framework applied (food use / Ph. Eur. 5.1.8 category). Buyer acceptance limits are TAMC ≤ X CFU/g, TYMC ≤ Y CFU/g, E. coli [absence in 1 g / ≤ Z CFU/g], Salmonella absent in 25 g, unless otherwise agreed in writing. If the material has been decontaminated, supplier shall declare the method (steam / irradiation) and, where irradiation is used, provide the mandatory declaration; ethylene-oxide-treated material is not accepted and an EO/2-CE residue statement shall be provided on request."

This protects both sides: the buyer gets a testable release gate, and the supplier avoids surprise rejection after production. It also stops regulatory and marketing language from mixing. ISO 22000 supports food-safety management, ISO 9001 supports quality-management discipline, and ISO 27001 protects the confidentiality of buyer specifications — but none of them is a microbial result for a specific lot. If species identity and grade also matter to your programme, align this with the identity and testing controls in the chamomile bulk sourcing guide and confirm scope on the Arovela certifications page.

Frequently asked questions

Why does Ph. Eur. allow such a high total count for herbal teas?

Because Category A material is intended to be prepared with boiling water. The infusion or decoction step acts as a consumer-side thermal treatment, so the pharmacopoeia tolerates a TAMC up to 10⁷ CFU/g on the dry drug while still requiring Salmonella absence in 25 g and an E. coli ceiling. Extract-grade Category B material is not brewed that way, so its TAMC ceiling drops to 10⁴ CFU/g and it adds a bile-tolerant Gram-negative limit. Naming the category is therefore essential before comparing any two COAs.

Is a high total aerobic count automatically a rejection?

Not by itself. Untreated dried herbs naturally carry high counts, and both the pharmacopoeia and food-industry practice recognise this. The decisive tests are usually the pathogen and indicator results — Salmonella absent in 25 g, E. coli within its limit, and yeast/mould below the agreed ceiling. A lot can show a high TAMC and still be acceptable for a boil-water tea, while the same count could fail an extract or ready-to-eat specification. Judge the result against the correct framework and intended use, not against a single remembered number.

Should buyers require steam-treated or untreated botanicals?

It depends on the finished product. Steam treatment reliably lowers microbial load and leaves no residue or label burden, which suits tight-limit or ready-to-eat programmes; the cost is potential aroma and colour loss that a premium tea or aromatherapy-adjacent buyer may reject. Untreated material can preserve sensory quality but must then meet the microbial spec on its own and may need a documented risk assessment for high natural counts. The RFQ should state whether treatment is required and, for EU imports, must exclude ethylene-oxide-fumigated material.

Source botanicals with a defensible microbial specification

If your team is buying dried herbs, botanicals or dried fruit from Turkey for the EU or Ukraine, Arovela can support lot-specific COA review, sampling documentation and export planning within its ISO 22000, ISO 9001 and ISO 27001 systems — without claiming certifications it does not hold. Start with a technical quote request, compare wholesale supply options, or review Arovela certifications before you finalise your TAMC, TYMC and pathogen limits.