An extract specification is a short document. Species, plant part, ratio, marker percentage, solvent, carrier, a handful of limits. Behind each of those lines sits a decision someone made in a factory, and the decisions interact: change the solvent and the marker percentage moves, change the drying method and the solubility changes, change the raw material's particle size and the extraction time and yield both change. Buyers who understand the process negotiate on the right variables. Buyers who do not end up comparing two numbers that were produced by different routes and are not comparable at all.
This guide walks through botanical extract manufacturing in the order it actually happens, from incoming plant material to a sealed drum of powder, and marks the points where a buyer can inspect, question or specify. It is deliberately process-focused: for the strategic comparison between the two dominant solvent routes, see our CO2 versus ethanol extraction guide, and for how ratios and markers are defined and misused, the extract standardisation guide goes into that specific question in detail.
Step 1: raw material preparation
Nothing downstream can rescue a bad input. This stage decides most of what the finished extract will be.
Identity. The plant must be confirmed as the declared species and the declared plant part. Root, leaf, flower, fruit and bark of the same plant have different constituent profiles, and substituting one for another is one of the oldest forms of adulteration in the trade. Confirmation is normally macroscopic and organoleptic, supported by chromatographic fingerprinting where the value or the risk justifies it.
Cleaning and sorting. Removal of stones, soil, metal, foreign plant material and damaged material. Weed contamination is not a cosmetic issue: for several herbal categories, co-harvested weeds are the origin of plant toxin contamination, which is why field discipline and cleaning both matter.
Drying and moisture control. Material arriving above target moisture will either be dried further or will deteriorate. Drying temperature is itself a quality decision, because heat drives off volatiles and can degrade heat-sensitive constituents.
Size reduction. Milling or cutting to a defined particle size. Smaller particles extract faster and more completely because of surface area, but excessive fines cause filtration problems, channelling in percolation columns and, at the extreme, a compacted bed that solvent flows around rather than through. A good manufacturer targets a distribution, not a maximum.
Batch homogenisation. Blending incoming lots to a consistent starting composition, which is how a manufacturer absorbs crop-year variation before it reaches your specification.
Buyer questions at this stage: what is the plant part, what is the origin region, what is the incoming acceptance specification, is the input tested for pesticide residues and heavy metals before extraction or only after, and what is the particle size target. That last question separates people who run a process from people who resell one.
Step 2: solvent choice
The solvent is the single most consequential decision in the whole sequence, because it determines which molecules leave the plant and which stay behind. Solvent selection follows polarity: polar solvents pull polar compounds, non-polar solvents pull lipophilic ones, and mixtures land in between.
- Water. Extracts polysaccharides, some glycosides, minerals, tannins and water-soluble acids. Cheap, entirely food-compatible, no residual solvent question. Downsides: low yield for lipophilic actives, a high microbial risk during extraction, and a large volume of water to evaporate afterwards, which means heat and cost.
- Ethanol and hydro-ethanolic mixtures. The industrial workhorse. Varying the ethanol-to-water ratio tunes polarity across a wide range, which is why specifications state a percentage — a 30% hydro-ethanolic extract and an 80% one from the same plant are genuinely different products. Ethanol is well accepted in food and supplement applications and is straightforward to recover and reuse.
- Glycerin. Used where an alcohol-free finished product is required. Lower extraction power for many constituents, and the glycerin remains in the product as a carrier.
- Supercritical CO2. Excellent for volatile and lipophilic fractions, leaves no solvent residue, runs at moderate temperature. Weaker for polar constituents unless a co-solvent such as ethanol is added. Capital-intensive, which is reflected in what it is used for.
- Other organic solvents. Acetone, hexane and others appear in some industrial processes. Their use in food applications is regulated in the EU, and where they are permitted, residue limits apply. If a solvent other than water, ethanol or CO2 appears in your supply chain, the residual solvent test is not optional — it is the specification.
Two things belong in the contract regardless of solvent: the named solvent and its concentration, and the residual solvent limit with the test method. An extract described only as "solvent: alcohol" is under-specified, and "solvent-free" as a marketing phrase means nothing without a residue figure.
| Route | Best suited to | Typical strengths | Trade-offs to plan for |
|---|---|---|---|
| Water / aqueous | Polysaccharides, water-soluble glycosides, traditional infusion profiles | No residual solvent, lowest regulatory friction, low input cost | Low yield on lipophilic actives; microbial control during extraction; high evaporation load |
| Hydro-ethanolic | The majority of polyphenols, flavonoids, most standardised supplement extracts | Tunable polarity, high yield, recoverable solvent, wide acceptance | Solvent recovery infrastructure needed; ethanol residue must be documented |
| High-strength ethanol | Resins, less polar constituents, some oleoresins | Strong extraction of moderately lipophilic fractions | Pulls chlorophyll and waxes; may require additional clean-up steps |
| Supercritical CO2 | Volatile aroma fractions, lipophilic actives, oleoresins | No solvent residue, low thermal load, selective | Weak on polar actives without co-solvent; capital cost; narrower applicability |
| Glycerin | Alcohol-free finished formats | Consumer-friendly declaration | Lower extraction efficiency; carrier stays in the product |
Step 3: the extraction itself
With solvent chosen, the mechanics vary:
Maceration soaks the milled material in solvent in a static vessel for hours or days, sometimes with agitation. Simple, gentle, solvent-hungry, and slow.
Percolation passes solvent through a packed bed of material in a column, so fresh solvent continuously meets partially exhausted plant matter. More efficient than maceration for the same solvent volume, and sensitive to bed packing — which is why particle size distribution matters so much.
Counter-current extraction moves solvent and solids in opposite directions through successive stages, so the most concentrated solvent meets the most exhausted material. It is the most solvent-efficient approach at industrial scale and the usual choice for continuous production.
Assisted methods — ultrasound and microwave — improve mass transfer and shorten extraction times. They are enhancements applied to one of the above rather than separate philosophies.
The parameters that a manufacturer controls, and that you can reasonably ask about, are temperature, time, solvent-to-solids ratio, number of extraction stages and agitation. Temperature is the one to probe: a hot, fast extraction gives good yield and can degrade the very constituents your marker measures. If your extract is defined by a heat-sensitive compound, ask what the maximum process temperature is, at every stage including drying.
Step 4: separation, clarification and concentration
The extraction produces a liquid loaded with dissolved constituents plus suspended solids.
Separation removes the spent plant material — pressing, filter presses, decanters, centrifuges. Yield is lost here if pressing is incomplete, and clarity is lost if filtration is inadequate.
Clarification removes fine particulate and, sometimes, unwanted classes of compound. Depth filtration, membrane filtration and, in some processes, resin treatment. Resin treatment is worth understanding because it can selectively enrich a class of compounds, and an extract enriched on resin is a different product from a simple concentrate even if both carry the same species name.
Concentration removes solvent, usually by evaporation under reduced pressure. Vacuum lowers the boiling point, which protects thermolabile constituents; this is where the difference between a careful and a careless manufacturer is often invisible on paper but obvious in the finished product's colour and aroma. Falling-film and wiped-film evaporators reduce residence time at temperature, which matters for sensitive material.
Solvent recovery. Recovered ethanol is normally redistilled and reused. Ask whether recovered solvent is tested before reuse, because carryover from a previous product is a real cross-contamination pathway in a multi-product plant.
Step 5: standardisation
Standardisation is the step that turns a concentrate into a specified commercial product, and it is where the most buyer confusion lives.
There are two distinct ways an extract is described:
- By ratio (drug-to-extract ratio, DER). For example 10:1 — ten parts of dried raw material yielded one part of finished extract. A ratio describes a process, not a composition. It is meaningful only when the raw material quality, the solvent and whether carriers were added are all stated, because adding carrier to a native extract changes the arithmetic without changing the plant chemistry.
- By marker content. For example a stated percentage of a named constituent, measured by a named method against a named reference standard. This is more informative — provided you know whether the marker is an active marker, an analytical marker used only for identity and consistency, or a class measurement by a non-specific method.
The reason both descriptions can be gamed is that both can be adjusted after extraction: by blending stronger and weaker batches, by adding carrier, or in bad-faith cases by adding an isolated compound bought separately so that the assay reads correctly while the extract behind it is weak. Non-specific colorimetric assays are especially easy to satisfy this way. This is why the assay method belongs on the certificate of analysis and why a chromatographic fingerprint — the whole profile, not one peak — is the practical defence. The standardisation guide covers the marker taxonomy and the specification wording in depth, and our botanical CoA guide covers reading the resulting document.
Step 6: drying and carriers
Most bulk extracts ship as powders, which means the concentrate must be dried.
Spray drying atomises the concentrate into a hot air stream. Fast, continuous, economical, and the standard route. It almost always requires a carrier — maltodextrin, gum acacia, dextrin or similar — because most botanical concentrates are hygroscopic and sticky and will not form a free-flowing powder alone. Contact time with hot air is short, which limits thermal damage.
Vacuum belt drying dries under reduced pressure at lower temperature, giving a gentler thermal profile and often allowing lower or no carrier. Slower and more expensive.
Freeze drying is the gentlest and the most costly, reserved for genuinely heat-sensitive material.
Carriers are not adulteration when they are declared: they are a physical necessity for handling. They become a problem when they are undeclared, because carrier content directly dilutes the native extract. A specification should state the identity and the percentage of every excipient, plus flow agents such as silicon dioxide. Two powders with the same headline ratio and different carrier loads are different products, and the cheaper one is usually cheaper for exactly that reason.
Finishing steps — sieving to a mesh specification, blending to homogenise, metal detection, packing under controlled humidity into lined drums — determine how the material behaves in your own plant.
Step 7: quality control points across the process
| Stage | What is checked | Why a buyer should care |
|---|---|---|
| Incoming raw material | Species and plant part identity, moisture, foreign matter, pesticide and heavy metal screening | An out-of-spec input cannot be corrected downstream; see the heavy metals guide |
| After milling | Particle size distribution, fines fraction | Drives extraction consistency and bed behaviour |
| In-process extraction | Temperature, time, solvent ratio, stage count | The record that proves batches were made the same way |
| After concentration | Solids content, colour, in-process assay | Early detection of a weak or over-processed batch |
| Standardisation | Marker assay with named method, chromatographic fingerprint | Where potency claims are made or broken |
| After drying | Moisture, water activity, bulk density, particle size | Handling, caking and stability in your own process |
| Finished lot | Full CoA: identity, assay, residual solvent, heavy metals, pesticides, microbiology | Your release decision and your regulatory file |
| Stability | Assay and appearance over time under defined conditions | Whether the specification still holds at the end of your shelf life |
Microbiological control deserves a note of its own. Aqueous stages are nutrient-rich and warm; the process must control contamination at those points rather than relying on a treatment at the end. The parameters and their interpretation are covered in the botanical microbial limits guide.
Regulatory context a buyer should carry into the conversation
Two areas catch buyers late.
First, novel food status. An extract is not automatically permitted in the EU merely because the plant is familiar. Where a preparation has no significant history of consumption in the Union before May 1997 — and a novel extraction route or a markedly higher concentration can create exactly that situation — authorisation may be required under Regulation (EU) 2015/2283. Our novel food risk guide explains how to assess this before you commit to a formulation.
Second, the function the extract performs in your product. An extract added for taste or aroma may fall under the flavourings framework of Regulation (EC) No 1334/2008, while the same botanical used for a technological effect may be treated as an additive, and one used as a characterising ingredient is neither. The classification changes your labelling and your dossier. Settle it before scale-up, not after. For the wider EU supply picture for supplement and food brands, see our botanical extracts sourcing guide.
What to ask a manufacturer, and the answers that should worry you
- "Which plant part, from which region, and how is identity confirmed?" Hesitation here is disqualifying.
- "Which solvent, at what concentration, and what is the residual solvent limit and method?" "Natural solvent" is not an answer.
- "What is the maximum process temperature, including drying?" Relevant for any heat-sensitive marker.
- "Is this a native extract or does it contain carrier, and at what percentage?" The answer should be a number.
- "Which assay method and which reference standard support the marker percentage?" A non-specific method behind a high headline number is a flag.
- "Can I see a chromatographic fingerprint for a recent lot?" The best single anti-adulteration document available to you.
- "Is the plant multi-product, and how is cross-contamination and solvent reuse controlled?" Especially relevant if you have allergen constraints.
- "What stability data exists, under which conditions?" If none exists, you are the stability study.
Arovela supplies bulk botanicals and botanical extracts to buyers in the EU and Ukraine, operating within ISO 22000, ISO 9001 and ISO 27001 management systems, with processing in Sındırgı, Balıkesir and a distribution warehouse in Solingen, Germany. We work to lot-specific documentation — DER, marker, method, carrier declaration, residual solvent and contaminant testing — and we are equally willing to tell you when a specification you have been quoted elsewhere does not add up. That conversation costs nothing and has saved buyers a great deal.
FAQ
Is a higher drug-to-extract ratio always a stronger extract?
No. A ratio describes how much raw material went in relative to finished weight, and it can be inflated by using lower-quality raw material or by comparing extracts with different carrier loads. Two 10:1 extracts of the same plant can differ substantially in marker content. Compare on the marker percentage, the assay method and the declared carrier content together.
What residual solvent level should I accept?
That depends on the solvent and the market. Ethanol is treated as a low-toxicity solvent and is tolerated at relatively high residual levels by the major pharmacopoeias, whereas solvents in higher-risk classes carry far tighter limits. The important thing is that the specification names the solvent, states a numeric limit and names the test method, and that the certificate of analysis reports an actual figure rather than "complies".
Why do two suppliers quote very different lead times for the same extract?
Usually because one is picking a catalogue item off a shelf and the other is running a campaign to your specification. A custom standardisation involves raw material procurement, a production run, analytical method work and a release cycle. Ask which of the two you are being quoted, because it changes both the lead time and how reproducible the next batch will be.
Can an extract be made without any carrier?
Sometimes. Native, carrier-free powders exist, typically produced by gentler drying routes, but many botanical concentrates are too hygroscopic to form a stable free-flowing powder without one. If carrier-free matters to you, expect a different drying method, different handling behaviour and a different cost structure, and ask for a sample that has been stored for a while rather than a fresh one.
How do I verify an extract has not been spiked with an isolated compound?
A single marker assay cannot tell you. A chromatographic fingerprint of the whole profile can, because a genuine extract shows the constellation of related compounds that the plant produces, while a spiked one shows a dominant peak without its natural companions. Ask for the fingerprint of the lot you are buying and compare it against a reference profile.
Specifying an extract you can reorder
The test of an extract specification is not whether the first batch is good. It is whether the fifth batch, from a different crop year, is recognisably the same product. That comes from naming the plant part, the solvent, the process temperature ceiling, the marker with its method, and the carrier — and from asking for a fingerprint rather than a number.
Send us the species, the marker you are working to, the application and the annual volume, and request a quote. If the specification you have been given is ambiguous, we will tell you which lines need fixing before anyone quotes a figure.