Plant Beverage OEM: Tannin Astringency, Debittering and Flavour Balance
In formula review for plant beverage OEM projects, the sensory defect that sends samples back most often is not weak aroma but astringency and lingering bitterness. Astringency shows up as a tightening, drying sensation across the mouth and tongue, while lingering bitterness leaves a persistent bitter note at the back of the mouth after swallowing. Both are common in herbal infusions, fruit and vegetable juices, plant protein drinks and botanical blends, and both tend to be masked by sweetness at sample stage, only surfacing when a reduced-sugar or ambient shelf-stable version is developed. The complication is that debittering and de-astringency treatments also strip the characteristic flavour and colour of the raw material, so over-processing leaves the product flat. This article follows raw material, extraction, fining and formulation routes to explain where the sensations come from, how to reduce them, and how to verify the result without losing character.
1. Where astringency and bitterness originate
Astringency is a tactile rather than a gustatory sensation, produced when polyphenols bind salivary and mucosal proteins and generate a drying, puckering feel. Tannins, chlorogenic acids and flavonoids are the main contributors, and the sensation strengthens with content and degree of polymerisation. Bitterness comes mainly from alkaloids, saponins, bitter peptides, certain amino acids and some flavonoid glycosides, and the perceived intensity and persistence differ by source: saponins are usually detected at the front of the mouth and fade quickly, alkaloids linger, and bitter peptides, which appear in plant protein drinks, intensify as protein content rises. Bitterness can also be generated during processing, from aldehydes and ketones produced by lipid oxidation, from scorching during excessive heating, and from metallic aftertaste contributed by certain sweeteners or mineral salts. Identify the source before choosing a treatment route.
2. Control at the raw material stage
Most astringency and bitterness problems are decided long before the blend. Variety is the most direct variable, since polyphenol content can differ several-fold between cultivars of the same plant, and selecting a low-astringency variety is often cheaper than removing the sensation later. Ripeness matters as well: unripe fruit carries more tannin, while over-ripe material risks fermented off-notes. Plant part has a strong effect, since leaves and stems hold more saponin than fruit flesh and bitter compounds in root materials often concentrate in the cortex, so peeling, coring and segregating parts at intake deliver a significant reduction. Post-harvest handling helps too: blanching inactivates polyphenol oxidase and cuts enzyme-derived bitterness, while controlled fermentation or drying alters the polymerisation state of the polyphenols. Writing variety, plant part, ripeness and key markers into the purchasing specification is the cheapest control available.
3. Extraction conditions and enzymatic assistance
Extraction is the main route by which bitter and astringent compounds enter the system, and therefore the most precisely controllable step. Excessively high temperature or long time raises the yield of both polyphenols and saponins, so staged extraction is common in practice: a lower-temperature, short contact step to draw out aroma and soluble solids, followed by a warmer step for the remaining actives, with the two streams blended to balance yield against taste. Solid-to-liquid ratio shifts the extraction equilibrium, and too little water concentrates the harsh fractions. Enzymatic assistance is an effective lever. Pectinase breaks down pectin, raises juice yield and improves body; protease hydrolyses some bitter peptides and reduces the coarse mouthfeel of plant protein; tannase cleaves the ester bonds of tannins, cutting astringency markedly and reducing later haze. Enzyme type, dose, temperature and time need a bench trial per material, because overdosing brings off-flavours or excessive hydrolysis and a thin result.
4. Fining, adsorption and debittering treatments
Fining serves stability but can also carry out de-astringency duty. Among common agents, gelatin with tannin precipitates high molecular weight tannins, bentonite and silica gel adsorb protein and some polyphenols, activated carbon adsorbs powerfully but without selectivity and readily strips colour and aroma, and resins offer higher selectivity at greater cost and with tighter residue control. Chitosan and PVPP target polyphenols more specifically with limited colour impact, which makes them common choices for fruit and vegetable drinks. Two cautions apply. First, verify marker content after treatment, because adsorption may remove functional actives along with the bitterness. Second, use only food-grade agents and validate that no residue remains, with dose and contact time written into the process instruction rather than adjusted at the line.
5. Formulation and flavour balancing
Formulation does not mask a problem so much as bring the bitterness back into an acceptable band. Sweetness is the most direct lever, but it does not shorten a lingering bitter note and can even accentuate it at high levels, so a better approach coordinates sweetness with acidity and adds some fat-related body. Acidity lifts overall freshness and changes perceived bitterness, although excess acid increases astringency, so a balance point has to be found. Milk or plant-based fat reduces perceived bitterness through encapsulation and dispersion while improving smoothness, which is why it is widely used in plant protein drinks. Cyclodextrin inclusion works on selected bitter compounds and suits low-dose, well-defined targets. Small amounts of salt and umami compounds round out body so a reduced-sugar product does not taste hollow. Once the flavour system is set, evaluate it on the reduced-sugar version so that sweetness is not doing the work.
6. Sensory verification and scale-up
Astringency and bitterness are subjective, so a repeatable evaluation method is essential. Quantitative descriptive analysis with a trained panel, scoring astringency, initial bitterness, lingering bitterness, sweetness return and smoothness, works well when the scores are correlated against analytical marker data to identify which compounds really drive perception. Timing matters too: freshly produced, stored, ambient and chilled samples all need evaluation, because polymerisation and precipitation change mouthfeel over time. Scale-up risk concentrates in the precision of extraction and enzymatic time and temperature control, in fining contact time, and in shifting homogenisation conditions, all of which can drift away from bench results. During pilot work, run at least three replicates of the same formula to establish the spread of sensory scores, and set the commercial specification from that spread.
Related Products
Milk Thistle, Kudzu & Turmeric Beverage with Sea Buckthorn Puree & Honey | Herbal Oral Liquid OEM | Liquid Fruit & Vegetable Salad Fiber Drink | Dietary Fiber Meal Replacement Vegetable Juice Beverage OEM
About Guangzhou Mesen Health Biotechnology Co., Ltd
Guangzhou Mesen Health Biotechnology Co., Ltd is located in the Huashan Science and Technology Park, Huadu District, Guangzhou. The company has focused on OEM and ODM services for the big health industry for nearly 20 years, operating dedicated research, production, sales and management centres, and maintaining long-term strategic partnerships with well-known Chinese medicine colleges and research institutions. Capabilities span pressed candy, gummies, capsules, solid beverages, herbal teas, plant beverages, protein powders, dietary supplements and enzyme jelly - from formulation development through contract manufacturing. For details on minimum order quantities, sampling lead times and quotations, see our manufacturing advantages and product range, or contact us for a tailored proposal.
Further Reading
