A practical, buyer-first playbook to set defensible OTA specs for Indonesian green coffee in 2025. Covers the EU roasted-coffee limit, a green-bean target, realistic green-to-roasted conversion, sampling plans, lab methods and LOQ, contract clauses, and what to do if a certificate fails.
If you buy Indonesian green coffee for the EU, your OTA spec can make or break a shipment. We’ve spent years setting, testing, and defending Ochratoxin A limits for wet-hulled Sumatras and fully washed Javas and Balis. Here’s the buyer playbook we use in 2025 so you can ship confidently and sleep at night.
The quick answer: what OTA spec should I set in green coffee?
For EU-bound roasted coffee, the legal limit is 3 µg/kg. There’s no EU legal limit for green coffee beans. So importers use a green-bean spec that reliably delivers ≤3 µg/kg after roasting.
Our 2025 recommendation for Indonesian green coffee OTA limit
- Standard roasted-coffee programs: set green OTA ≤2.0 µg/kg (composite, LC-MS/MS, LOQ ≤0.2 µg/kg).
- High-risk wet-hulled origins (e.g., Mandheling/Lintong during humid months) or light roast profiles: tighten to ≤1.5–1.8 µg/kg.
- Instant/soluble inputs: aim ≤1.2–1.5 µg/kg because extraction can concentrate OTA and you must meet a stricter finished limit.
We don’t assume roasting cleans everything up. In our experience, it often doesn’t. Need help tailoring the number to your roast style or product mix? Contact us on whatsapp.
Is there an EU legal limit for OTA in green coffee beans?
No. EU rules apply to the food as placed on the market. For coffee, Regulation (EU) 2023/915 caps OTA in roasted coffee at 3 µg/kg. Instant coffee has its own specific limit. That’s why buyers back-calculate a green-bean spec instead of referencing a non-existent legal limit for green.
So what limit should you set for Indonesian green to meet the EU 3 µg/kg roasted limit?
For 2025, we set 2.0 µg/kg for most lots, then adjust by risk:
- Fully washed, well-dried Arabicas like Arabica Java Ijen Grade 1 Green Coffee Beans or Arabica Bali Kintamani Grade 1 Green Coffee Beans: ≤2.0 µg/kg is usually comfortable.
- Wet-hulled Sumatras like Sumatra Mandheling Green Coffee Beans or Sumatra Lintong Green Coffee Beans (Lintong Grade 1), especially during rainy harvests: specify ≤1.5–1.8 µg/kg. The process and climate raise OTA risk if drying and storage slip.
Does roasting save you? The conversion that actually works
How much does roasting change OTA levels?
Here’s the thing. OTA is heat-stable. We’ve seen reductions from 0% to about 30% depending on roast time/temperature, drum type, and bean chemistry. Light and medium roasts often show minimal change. Dark roasts can reduce OTA more, but not predictably enough to bank on.
What’s interesting is that moisture loss during roasting can offset any OTA destruction. If mass drops 16–18% and OTA falls only 10–20%, the concentration in the roasted beans might barely change. In some light roasts, it even increases slightly.
What conversion factor should I use?
For contracts and compliance calculations, we use a conservative green-to-roasted conversion of 1.0. In other words, assume the roasted concentration will be roughly the same as the green composite result.
- If you plan a darker roast, you can sometimes justify a 0.85–0.95 factor. But document it with side-by-side tests on your roaster.
- For instant/soluble, treat the process as a potential concentrator and set an even tighter green spec (≤1.2–1.5 µg/kg).
Practical takeaway: set your green spec so that even at a 1.0 conversion you’re inside 3 µg/kg. Don’t rely on the roast to fix an out-of-spec green lot.
Sampling that stands up in audits
How many bags should I sample in a 320‑bag lot?
OTA can be spotty. A shallow sampling plan creates false confidence. For a 19–20 MT container around 320 bags, we recommend:
- Incremental bag selection: 60 bags minimum for wet-hulled Indonesian Arabica. 30 bags minimum for fully washed Indonesian Arabica or Robusta in low-risk periods.
- Sample mass: take ~200 g from each selected bag using a trier at varying depths. Mix to form a composite.
- Composite size: 12 kg target composite for 60-bag plans, then reduce to a 1–2 kg lab sample via riffle or rotary splitter.
If you need to defend it in an audit, we cite alignment to mycotoxin sampling principles in EU guidance and ISO approaches to heterogeneous bulk goods. The key is enough increments and a proper composite.
How should the composite be prepared for the lab?
- Grind the composite to a uniform particle size to reduce subsample variance.
- Split into triplicates: one for the lab, one retain at origin, one retain at destination.
- Seal and label with tamper-evident tape and full traceability (lot ID, bag range, sampler, date).
Practical takeaway: for Indonesian wet-hulled lots, 60 incremental samples beats 10 every time. It costs a bit more. It saves you from surprises.
The lab method, LOQ, and what “<LOQ” really means
Which OTA testing method should I specify?
- Preferred: LC‑MS/MS with solid-phase or immunoaffinity cleanup. Best selectivity and lowest LOQs, especially in complex matrices.
- Acceptable: HPLC‑FLD with immunoaffinity column cleanup. Widely available and reliable when properly validated.
In Indonesia, use ISO/IEC 17025 (KAN) accredited labs. We routinely work with reputable third-party labs for pre-shipment OTA certificates.
What LOQ should be on the certificate?
Set LOQ ≤0.2 µg/kg for green coffee. Many labs can achieve 0.1–0.2 µg/kg. This matters for two reasons:
- You need resolution when your spec is 1.5–2.0 µg/kg.
- You need to interpret “<LOQ.” We treat results reported as “<LOQ” as compliant and effectively zero for acceptance decisions. For trending, we record LOQ/2.
Practical takeaway: state the method and LOQ explicitly in contracts and POs. Otherwise you’ll get beautiful certificates that aren’t decision-grade.
Contract language you can copy
What should my purchase contract say about OTA and retests?
Here’s a concise clause our buyers have used successfully:
- Maximum OTA in green coffee (composite): ≤2.0 µg/kg for standard EU-roast programs. Method: LC‑MS/MS or HPLC‑FLD with IAC cleanup. LOQ ≤0.2 µg/kg. Lab: ISO/IEC 17025 accredited.
- Sampling: Composite from minimum 60 bags for wet-hulled lots, 30 bags for fully washed lots. ~200 g per bag, composite ground and split. Buyer or appointed surveyor may witness.
- Certificates: Pre-shipment OTA certificate required. Results reported as “<LOQ” deemed compliant.
- Retest plan: If 2.01–2.5 µg/kg, perform two additional independent composites from retained samples. Accept if the mean of three results ≤2.0 µg/kg and no single value >2.5 µg/kg. Above 2.5 µg/kg triggers non-conformance.
- Remedies: At buyer’s option, supplier sorts/reprocesses at origin with new certificate, price adjustment, or lot rejection at seller’s cost.
Tweak the numeric targets for higher-risk seasons or instant-coffee inputs.
What do I do if a pre-shipment OTA test fails?
- Verify the method and LOQ. Ask for chromatograms and QC recoveries. We’ve seen “fails” disappear after a proper cleanup step or corrected matrix spiking.
- Recompose and retest from sealed retains. If you only tested a tiny subsample of the composite, you might be chasing noise.
- If still high, triage the lot. Remove screen fines, broken beans, and high-defect bags. With Indonesian wet-hulled coffee, targeted resorting often drops OTA by 10–30%.
- Re‑dry if moisture is >12.5%. OTA can rise in transit if beans are borderline and containers sweat.
- If after rework you can’t meet the green spec, divert to non‑EU or blend under a validated plan that keeps the roasted outcome ≤3 µg/kg. Document everything.
Practical extras that lower OTA risk in Indonesian supply
A few things we’ve found that move the needle:
- Aim for moisture 10.5–12.0% at stuffing. Pair with container liners and adequate desiccants. OTA can grow in-transit if condensation develops.
- Remove fines and broken beans before sampling. This makes your composite reflect the exported lot, not mill dust.
- Avoid rebagging after sampling. If rebagging is necessary, resample. Mixing can change the profile.
- For hand-sorted peaberries like our Sumatra Super Peaberry Green Coffee Beans, the defect load tends to be lower, and we see more stable OTA behavior.
If you prefer to start with cleaner, fully washed profiles for EU programs, consider Arabica Java Ijen Grade 1 Green Coffee Beans or Arabica Bali Kintamani Grade 1 Green Coffee Beans. We also ship wet-hulled classics like Sumatra Mandheling Green Coffee Beans under a tightened OTA spec and enhanced sampling.
Key takeaways you can use today
- There’s no EU limit for green beans. The roasted limit is 3 µg/kg. Build your green spec to guarantee that outcome.
- Use a conservative green-to-roasted conversion of 1.0 unless you’ve validated your roast.
- For Indonesian lots, write 2.0 µg/kg into your contract, with 1.5–1.8 µg/kg for higher-risk wet-hulled seasons. Instant inputs should be ≤1.2–1.5 µg/kg.
- Sample enough bags. Sixty increments for wet-hulled containers is the difference between luck and confidence.
- State the method and LOQ. LC‑MS/MS preferred. LOQ ≤0.2 µg/kg. “<LOQ” counts as compliant.
Questions about your current lot, roast profile, or contract wording? View our products or reach out and we’ll share what’s working for EU buyers this season.