Pesticide Residues in Botanicals: MRLs, Testing & Global Rules

Botanicals are agricultural products, and agricultural products carry the chemistry of the field they grew in. Pesticide residues — insecticides, fungicides, herbicides and their breakdown products — can travel from the growing crop through drying, milling and extraction and into the material a formulator finally receives. The questions that matter are not whether residues are theoretically possible (they almost always are) but which compounds were screened for, at what sensitivity, against whose limits, and on which form of the material. Reading a residue report well means understanding how residues behave during processing and how the rules differ across the markets you sell into.

Why pesticide residues matter in extracts

Residues matter for two distinct reasons, and conflating them leads to weak specifications. The first is safety: certain compounds — organochlorines, some organophosphates and a handful of fungicides — are toxicologically significant, and regulators set limits precisely to cap consumer exposure. The second is compliance and market access: even a residue with no realistic safety concern at the level found can render a lot non-compliant if it exceeds the maximum residue limit (MRL) of the destination market, blocking import or triggering a recall. A botanical extract can be perfectly safe to ingest and still be illegal to sell in a given country because a single residue sits above that country's limit.

For extracts specifically, the stakes are higher than for raw herbs, because extraction can change the residue picture in ways that are easy to miss. A buyer who tests the raw herb, sees a pass, and assumes the extract inherits that pass is making an assumption that the chemistry does not support.

How residues concentrate during extraction

This is the point most often missed in botanical sourcing, so it is worth stating plainly: residues can concentrate in the extract relative to the raw herb. An extract is, by definition, a concentrated fraction of the plant. If a 5:1 extract is made — five kilograms of dried herb yielding one kilogram of extract — every compound that survives the process and reports into the extract is, in principle, concentrated by up to that drug-to-extract ratio. A residue present at a comfortable level in the leaf can land several-fold higher in the powder.

Whether a given residue concentrates, dilutes or disappears depends on its physical chemistry relative to the extraction solvent and conditions. The governing factor is solubility and partitioning.

Lipophilic residues and solvent affinity

Many older and persistent pesticides — organochlorines in particular — are lipophilic (fat-loving) and non-polar. In an ethanolic or hydroalcoholic extraction they partition readily into the solvent and report strongly into the extract. In an oily or supercritical-CO2 extract they can concentrate dramatically, because the residue and the target actives share the same solubility behaviour. The same affinity that pulls the desired constituents out of the plant pulls the lipophilic residue out with them.

Polar residues and aqueous behaviour

Polar, water-soluble pesticides behave differently. In a predominantly aqueous extraction they may follow the water phase and concentrate in a water-based extract; in a non-polar process they may be left behind in the spent plant material (the marc). The result is that the residue profile of an extract is not a scaled copy of the raw herb's profile — it is a re-weighted profile shaped by the solvent system.

MRL frameworks differ by market

There is no single global pesticide limit. Each major market maintains its own framework, set by a different authority, with different compound coverage and different numerical limits — and a value can vary by orders of magnitude between them. A lot built to one market's rules may breach another's, so the destination governs the specification.

European Union — the EU MRL database

The EU operates a centralised, legally binding MRL framework under Regulation (EC) No 396/2005, published in the searchable EU Pesticides MRL database. It sets a specific MRL for each pesticide–commodity pair, expressed in mg/kg (ppm). Where no specific MRL exists, a default of 0.01 mg/kg applies — effectively the limit of quantification — which is highly demanding for botanicals and a frequent cause of rejection. The EU framework is generally the most stringent and the most consequential for exporters of herbal materials.

United States — EPA tolerances

In the US, residue limits are called tolerances and are set by the Environmental Protection Agency (EPA) under the Federal Food, Drug, and Cosmetic Act, then enforced by the FDA. Tolerances are also commodity-specific and expressed in ppm. Critically, where no tolerance has been established for a pesticide–commodity pair, the default position is zero — any detectable residue can make the food adulterated. EPA coverage of specialty botanical crops is uneven, so the absence of an established tolerance is itself a compliance trap.

Pharmacopoeial limits — USP and EP

For materials traded as pharmacopoeial-grade botanicals, the limits come not from food regulators but from the pharmacopoeias. The European Pharmacopoeia (EP) general chapter on pesticide residues (2.8.13) and the corresponding US Pharmacopeia (USP) chapters list specified residues with individual limits and a calculation approach that can also bound total residues by reference to daily intake. These pharmacopoeial limits apply to herbal drugs and their preparations and are the relevant benchmark when an extract is sold against a USP or EP monograph rather than as a food ingredient.

Japan — the positive list system

Japan operates a positive list system: every pesticide is regulated. Compounds with a specific MRL are held to that MRL; every other compound is subject to a uniform default limit of 0.01 ppm. Because nothing is unregulated, the positive list is comprehensive and unforgiving for botanicals, where many residues have no crop-specific value and therefore fall to the 0.01 ppm floor.

Multi-residue testing: GC-MS/MS and LC-MS/MS

Modern pesticide screening is multi-residue: a single analytical run targets hundreds of compounds at once, rather than testing for them one by one. Two complementary techniques carry the work, and a credible screen uses both because no single instrument sees every compound class.

• GC-MS/MS (gas chromatography–tandem mass spectrometry) — resolves volatile and thermally stable, typically non-polar residues such as organochlorines, organophosphates, pyrethroids and many fungicides.
• LC-MS/MS (liquid chromatography–tandem mass spectrometry) — handles polar, thermally labile and non-volatile residues — carbamates, many modern systemics and neonicotinoids — that do not survive a GC inlet.

Sample preparation usually follows a QuEChERS-style extraction (Quick, Easy, Cheap, Effective, Rugged and Safe) before injection. The tandem-MS (MS/MS) step provides the specificity that complex botanical matrices demand: by monitoring characteristic precursor-to-product ion transitions, the method distinguishes a true residue from the dense background of natural plant constituents that would otherwise generate false signals. A panel that screens several hundred residues across both platforms is the practical standard for a botanical extract.

Why coverage and sensitivity both matter

Two numbers define the quality of a screen: how many compounds it covers, and how low it can reliably see each one (the limit of quantification, LOQ). A 600-compound panel that only reaches an LOQ of 0.05 mg/kg cannot demonstrate compliance with an EU 0.01 mg/kg default — the method is blind below the limit it is meant to police. For the most demanding markets, the LOQ must sit at or below the applicable MRL, or the report cannot actually prove a pass.

Raw-herb pass versus extract pass

This distinction follows directly from the concentration effect, and it is where many residue programmes quietly fail. A certificate showing that the incoming raw herb met its MRLs is necessary but not sufficient. The extract is a different material with a re-weighted, potentially concentrated residue profile, and it must be evaluated against the MRL as the form actually sold.

1. 1A raw-herb residue report characterises the botanical before extraction — useful for sourcing and supplier control, but not the finished-good specification.
2. 2An extract residue report characterises the material as shipped — the only report that demonstrates the finished extract meets the limit.
3. 3Where a residue concentrates through a high drug-to-extract ratio, only the extract report can catch a value that the raw-herb report passed.
4. 4Where MRLs are written for the herb and the product is an extract, the limit must be interpreted for the form sold — concentration factor and intended use both inform that reading.

Sampling and how to specify

A residue result is only as representative as the sample behind it. Pesticide application and residue distribution can be uneven across a field, a harvest and even a single lot, so a non-representative sample can both miss a hotspot and overstate a clean lot. Sound sampling — drawing incremental samples across the lot and compositing them to a homogeneous laboratory sample, following an established plan such as the relevant Codex or pharmacopoeial sampling guidance — is the foundation of a trustworthy number.

When writing a residue clause into a specification, make it unambiguous on each of the following points.

• Form tested — state that testing is on the finished extract, not only the raw herb.
• Markets — name every destination market whose MRLs must be met (e.g. EU MRL database, US/EPA, USP/EP, Japan positive list).
• Panel scope — specify a multi-residue panel covering both GC-MS/MS- and LC-MS/MS-amenable compounds, with a stated number of analytes.
• Sensitivity — require an LOQ at or below the lowest applicable MRL (often the EU 0.01 mg/kg default).
• Acceptance basis — state that each residue must meet the applicable MRL for the named market(s), per lot.
• Frequency and traceability — per-lot testing with a COA that names the method, the panel, the LOQ and the limits applied.

Organic versus conventional material

Certified organic material is grown without synthetic pesticides, which lowers — but does not eliminate — residue risk. Spray drift from neighbouring fields, contaminated irrigation water, persistent legacy compounds in the soil, and adulteration can all introduce residues into nominally organic botanicals. Organic certification is a process and provenance claim; it is not a residue test result, and the two should not be substituted for one another.

Conventional material carries a higher prior probability of residues and makes residue testing of the finished extract essential rather than optional. For both grades, the defensible position is the same: verify by analysis on the extract. Organic status reduces expected risk and supports a leaner panel or frequency; it does not replace the test, because the concentration effect operates regardless of how the herb was grown.

Questions to ask your supplier

Before accepting a residue claim into your own specification, put it through these questions in order. The answers separate a finished-extract demonstration from a raw-material assumption.

1. 1Was residue testing performed on the finished extract, or only on the incoming raw herb?
2. 2What is the drug-to-extract ratio, and has the concentration effect on residues been considered for this process and solvent?
3. 3Which markets' limits does the lot meet — EU MRL database, US/EPA tolerances, USP/EP, Japan positive list — and can you confirm each?
4. 4How many compounds does the panel cover, and does it run on both GC-MS/MS and LC-MS/MS?
5. 5What is the LOQ for each compound, and does it sit at or below the lowest applicable MRL (e.g. the EU 0.01 mg/kg default)?
6. 6How was the lot sampled, and does the sampling plan follow a recognised protocol?
7. 7For organic material, is there finished-extract residue data, or is the claim resting on certification alone?
8. 8Is residue testing performed per lot, and does the COA state the method, panel, LOQ and the limits applied?