In still wine processing, the addition of sorbic acid or potassium sorbate is assigned specifically to the interval after confirmed malolactic inactivation and before final filtration because sorbic acid acts as a yeast reproduction inhibitor rather than a sterilant. The undissociated acid fraction penetrates yeast cells and dissipates the proton motive force, whereas the dissociated form remains predominantly extracellular. The pKa of sorbic acid is 4.76 at 25°C; therefore, at wine pH 3.2, 3.5, and 3.8, the undissociated acid fraction is approximately 97.3%, 94.8%, and 90.1%, respectively. This pH-dependent activity does not extend to lactic acid bacteria, acetic acid bacteria, or malolactic bacteria at legal dose rates, and sorbic acid is not a substitute for molecular SO2 or terminal microbial filtration. The legal maxima in major wine-producing jurisdictions are 200 mg/L expressed as sorbic acid under Regulation (EC) No 606/2009 and the OIV Code of Oenological Practices, and 300 mg/L under U.S. TTB 27 CFR 24.246. Because sorbic acid does not remove existing viable yeast biomass, the process sequence places the addition after malolactic inactivation but before the final membrane barrier so that any contaminant introduced during preservative dosing is retained downstream. This sequence also permits dissolution, homogenization, and analytical verification before bottling, but it imposes specific constraints related to lactic acid bacteria metabolism, potassium bitartrate stability, membrane fouling, and residual fining agents.
The primary microbiological boundary for this addition point is the metabolic capacity of residual lactic acid bacteria to transform sorbic acid into 2-ethoxyhexa-3,5-diene, the compound associated with geranium-like off-odor. This transformation occurs when viable Oenococcus oeni or other lactic acid bacteria remain active after sorbate addition. Malolactic inactivation must therefore be verified before sorbic acid is introduced. Verification typically requires enzymatic malic acid analysis demonstrating that malic acid has reached a validated endpoint, commonly 0.1 g/L or below, although the exact endpoint depends on initial malic acid concentration and wine style. Absence of viable lactic acid bacteria should be confirmed by plate count, qPCR, or equivalent cell viability testing rather than assumed from malic acid depletion alone. Production-scale batch records should document lot-specific malolactic inoculation date, fermentation temperature, and malic acid depletion curve. Batch-to-batch variation in malolactic completion is a known processing bottleneck; sorbic acid addition to an incompletely inactivated population remains a principal cause of geranium taint-related rejections. The addition of sorbic acid must therefore be refused or postponed when malolactic activity is incomplete, when viable lactic acid bacteria are detected, or when the wine has not been maintained under sufficient molecular SO2, lysozyme, or other validated inhibitory conditions. This incompatibility is absolute: sorbic acid must never be combined with an active malolactic starter or introduced into a tank with residual malolactic activity. Furthermore, sorbic acid is not effective against acetic acid bacteria and does not provide bacterial sanitation; the wine must already be low in viable bacteria through upstream processing and SO2 management. For still wines with residual sugar, the preservative function of sorbic acid is directed against post-filtration yeast growth, and its efficacy should be interpreted as yeaststatic rather than yeastcidal at legal concentrations. Some preservative-resistant yeasts, including Zygosaccharomyces bailii, may tolerate sorbic acid at the permitted maximum; therefore low-temperature storage and terminal filtration remain essential control points even when sorbic acid is present at 200 mg/L.
Sorbic acid can also undergo slow non-enzymatic esterification with ethanol under acidic wine conditions to form ethyl sorbate. This compound is distinct from the geranium taint compound and is generally considered less sensorially active at legal doses, although published data for this specific configuration in long-aged still wines is limited. The more significant sensory risk is microbial transformation by residual lactic acid bacteria, which can occur if malolactic bacteria recover after bottling. For this reason, the addition point between malolactic inactivation and filtration must be supported by post-bottling microbial stability protocols that include free SO2 maintenance and storage temperature control. At pH 3.2–3.5, a free SO2 reserve of 20–40 mg/L may be required to achieve approximately 0.8 mg/L molecular SO2, which is the commonly applied antimicrobial threshold for still wine. Sorbic acid is not a substitute for this molecular SO2 reserve, and wines with low free SO2 remain susceptible to bacterial spoilage even when sorbic acid is present at the legal maximum.
When potassium sorbate is selected as the delivery salt, the preparation vessel should be sized for complete dissolution before transfer into the receiving tank. The stoichiometric conversion is based on molecular masses of 112.13 g/mol for sorbic acid and 150.22 g/mol for potassium sorbate; 1.0 g of sorbic acid is equivalent to 1.34 g of potassium sorbate, while 1.0 g of potassium sorbate supplies 0.746 g of sorbic acid. Sorbic acid has a water solubility of approximately 1.6 g/L at 20°C, but potassium sorbate has a solubility of approximately 58.5 g/100 mL at 25°C, making the potassium salt preferred for aqueous stock solutions. For a 10,000 L still wine tank targeted at the EU maximum of 200 mg/L sorbic acid, the required mass is 2.0 kg sorbic acid or 2.68 kg potassium sorbate. A stock solution prepared at 20% w/v potassium sorbate would require 13.4 L of water. If sorbic acid itself is used, dissolution in ethanol or warm water and injection through a Venturi or static mixer is required because direct addition of acid flakes to cold wine may leave undissolved material. The equimolar addition of sorbic acid increases titratable acidity by approximately 0.13 g/L expressed as tartaric acid, whereas potassium sorbate adds approximately 70 mg/L potassium at the 200 mg/L sorbic acid equivalent dose. That potassium input is stoichiometrically equivalent to approximately 0.34 g/L potassium bitartrate, although actual precipitation depends on existing tartrate, ethanol, and temperature conditions. This dichotomy represents a process conflict: low-potassium wines may tolerate potassium sorbate, but wines already near potassium bitartrate saturation should be re-evaluated by cold stability testing after addition. In high-bitartrate still wines, sorbic acid may be preferred to avoid additional potassium, but the associated titratable acidity increase must then be considered against the sensory and legal profile of the wine.
| Parameter | Sorbic acid | Potassium sorbate |
|---|---|---|
| Molecular weight | 112.13 g/mol | 150.22 g/mol |
| Sorbic acid equivalence | 1.000 g/g | 0.746 g/g |
| Water solubility | 1.6 g/L at 20°C | 58.5 g/100 mL at 25°C |
| Dose for 10,000 L at 200 mg/L sorbic acid | 2.0 kg | 2.68 kg |
| Main process load at 200 mg/L sorbic acid equivalent | +0.13 g/L titratable acidity as tartaric | +70 mg/L potassium, equivalent to 0.34 g/L potassium bitartrate stoichiometric load |
For production-scale addition, the dissolved preservative should be introduced through a dosing port upstream of a static mixer or into a recirculation loop with a centrifugal pump delivering at least 1.5–2.5 tank turnovers per outlet sample. A 50–100 µm inline strainer downstream of the dosing point removes undissolved granules and prevents fouling of downstream plate-and-frame or cartridge filters. In tanks with bottom-mounted agitators, an impeller speed of 300–500 rpm may be used for 20–30 min, but the actual mixing time must be verified by sorbic acid concentration sampled from top, middle, and bottom sample valves; deviations greater than ±5% of target indicate incomplete turnover. Batch-to-batch variation in filterability after sorbate addition is frequently caused by cooling of concentrated potassium sorbate solutions below 15°C, which can cause recrystallization in transfer lines. Jacketed dissolution vessels maintained at 20–25°C and short transfer lines with positive displacement pumps prevent this failure mode. Concentrated potassium sorbate solution can also have an alkaline local pH, so injection into a high-flow recirculation loop rather than directly into a stagnant tank is required to prevent local pH excursion.
After sorbate addition, the final filtration step should be treated as a particle-removal and yeast-removal operation, not as a sorbate-retention step. In a fully dissolved state at legal concentrations, sorbic acid and potassium sorbate pass through 0.45 µm polyethersulfone, polyvinylidene fluoride, and nylon membranes without measurable sieving retention. Membrane selection is therefore driven by bacterial retention validation such as ASTM F838-20, with bubble point or forward-flow integrity tests performed according to the membrane manufacturer’s specification for 0.45 µm membrane cartridges. However, if sorbate addition is performed as a slurry or if recrystallization occurs, the resulting particles accumulate on 1.0–5.0 µm depth pre-filters and reduce differential pressure. Production-scale crossflow units and plate-and-frame stacks should be protected by a 0.45 µm final membrane with upstream 1.0 µm and 0.65 µm prefiltration cartridges to avoid premature plugging. Filter flux decay above 25–30% over the first 30 min typically indicates incomplete dissolution or potassium bitartrate precipitation, not sorbate-membrane incompatibility. Because sorbic acid does not provide sterilization, the final membrane must retain viable yeast and bacteria. In still wine, a 0.45 µm membrane is considered a terminal microbial barrier when integrity testing passes, while a 0.65 µm membrane may be insufficient for some bacterial contaminants. The addition point before final filtration also reduces post-filtration recontamination risk from dosing pumps and connections, because any inadvertent microbial challenge introduced during preservative addition is removed by the downstream membrane. Published data for sorbate adsorption to PVDF or PES membrane surfaces under wine pH conditions is limited, but at 200 mg/L no measurable retention is reported when the preservative is fully dissolved.
Wines containing residual molecular SO2, lysozyme, or ascorbic acid impose specific constraints on sorbate addition because these constituents alter microbial ecology, redox buffering, and protein haze behavior. Molecular SO2 remains the primary bacterial inhibitor, and sorbic acid should not be relied upon for bacterial suppression when molecular SO2 falls below 0.8 mg/L for the given pH. For typical still wine pH values of 3.2–3.5, the corresponding free SO2 requirement to reach 0.8 mg/L molecular SO2 ranges from approximately 20 mg/L to 40 mg/L; at pH 3.8, the required free SO2 rises to approximately 77 mg/L. Lysozyme, if used for malolactic inactivation, is a positively charged protein that can remain soluble and later contribute to protein haze if not removed by bentonite fining. Sorbic acid addition should therefore follow bentonite fining and protein stability verification; otherwise the downstream filtration step may be insufficient to remove heat-unstable lysozyme complexes from the finished wine. Ascorbic acid does not directly complex with sorbic acid, but it alters redox potential and can exacerbate oxidative browning when dosed without adequate SO2. Sorbate addition does not compensate for ascorbic acid redox behavior, and wines receiving ascorbic acid at 50–100 mg/L must be managed through free SO2 measurement, dissolved oxygen control, and pre-bottling oxidation checks rather than through sorbic acid concentration alone. The presence of residual ascorbic acid does not contraindicate sorbate use, but it increases the need for post-addition analytical verification because ascorbic acid can interfere with some older spectrophotometric preservative assays if not separated chromatographically.
After the final filtration step, quantification of sorbic acid requires a representative sample withdrawn from a recirculation loop after at least two full tank turnovers or from post-filter bottling line take-offs. The standard liquid chromatographic determination uses reversed-phase C18 columns with UV detection at 255 nm; the UV maximum for sorbic acid is near 254–256 nm, with quantification against external standards prepared in a wine matrix to avoid recovery shifts. Analytical tolerance and adjustment should be built into the dosing calculation. For a target of 180–200 mg/L, the initial bulk addition should not exceed 190 mg/L to allow homogenization variability and avoid exceeding the legal limit in any final package. If potassium sorbate is used, the measured sorbic acid concentration must be expressed as sorbic acid for compliance, and the final release value should be verified in the finished wine after final filtration because some pre-filters may retain precipitated or undissolved material. The compliance matrix in Table 2 summarizes the applicable maxima and analytical approaches.
| Jurisdiction or standard | Maximum sorbic acid | Typical analytical method |
|---|---|---|
| European Union, Regulation (EC) No 606/2009 | 200 mg/L | HPLC-DAD at 255 nm |
| OIV Code of Oenological Practices | 200 mg/L | OIV Compendium HPLC method |
| United States, 27 CFR 24.246 | 300 mg/L | HPLC-UV at 255 nm |
Operational boundaries include the incompatibility of sorbic acid with active malolactic bacteria, the requirement for complete dissolution before downstream membrane stacks, and the need to re-test bitartrate cold stability when potassium sorbate is used at the legal maximum. Sorbic acid is not effective against acetic acid bacteria, is not a substitute for molecular SO2, and does not remove existing microbial biomass. The sorbate addition point between malolactic inactivation and filtration is valid only when the wine is already low in viable yeast and bacteria through upstream processing. Published data for the interaction between sorbic acid and some crossflow membrane materials at high pH and low temperature are limited; for prolonged contact with pH above 3.8, the undissociated fraction falls below 90%, and antimicrobial efficacy should be revalidated with challenge testing rather than assumed from legal limits alone.