Geranium Taint — Sorbate + Lactic Bacteria Reaction (2-Ethoxyhexa-3,5-diene)

Geranium taint is a wine fault characterised by a powerful, disagreeable aroma of crushed geranium leaves. It results from the microbial degradation of sorbic acid, added to wine as potassium sorbate, by certain lactic acid bacteria. The culprit compound, 2-ethoxyhexa-3,5-diene, was first identified and reported by Crowell and Guymon in 1975 at UC Davis. This is strictly a sorbate-related microbial fault, distinct from cork taint, volatile acidity, or sulfur-related reductive faults.

• Primary compound: 2-ethoxyhexa-3,5-diene, a C6 volatile ether derived from the acid-catalysed isomerisation and ethanolysis of sorbic alcohol
• Sensory threshold less than 1 ng/L, confirmed by Chisholm and Samuels (1992) using gas chromatography-olfactometry, making it extremely potent
• Distinct from mercaptan faults (struck match, rubber) and volatile acidity (vinegar); geranium taint has a uniquely floral-solvent character
• Because ethanol is required for the final ether-forming reaction, the taint does not develop in grape must, only in finished wine

Potassium sorbate is added to inhibit yeast re-fermentation, but sorbic acid is not a bactericide and does not suppress lactic acid bacteria. Certain LAB, particularly Oenococcus oeni strains, can metabolise sorbic acid to sorbic alcohol (2,4-hexadien-1-ol). Under the acidic conditions of wine, this alcohol isomerises to 3,5-hexadien-2-ol, which then reacts with ethanol to form 2-ethoxyhexa-3,5-diene. This multi-step pathway was characterised by Crowell and Guymon (1975) and further investigated by Edinger and Splittstoesser (1986).

• Step 1: LAB reduce sorbic acid to sorbic alcohol (2,4-hexadien-1-ol) via reductive metabolism
• Step 2: Acid-catalysed isomerisation converts 2,4-hexadien-1-ol to 3,5-hexadien-2-ol in the wine matrix
• Step 3: 3,5-hexadien-2-ol reacts with ethanol to form 2-ethoxyhexa-3,5-diene, the true geranium taint compound
• Edinger and Splittstoesser (1986) demonstrated that Leuconostoc strains (the group containing Oenococcus oeni) produced sorbic alcohol; Lactobacillus and Pediococcus strains in their study did not

Geranium taint is not an inevitable consequence of sorbate use, but specific conditions must align for LAB to survive in sufficient numbers to degrade sorbic acid. The greatest risk arises when potassium sorbate is added to wines that still contain active LAB populations, particularly after malolactic fermentation. pH, free SO2 level, temperature, and pre-bottling microbial load all influence whether the taint develops. Maintaining adequate free SO2 alongside sorbate is essential, since sorbic acid alone does not suppress bacteria.

• Highest-risk scenario: sorbate added to a wine that has undergone MLF and still carries residual LAB without adequate free SO2 suppression
• pH above 3.5 reduces the efficacy of SO2 against LAB, increasing the chance of bacterial survival alongside sorbate
• High free SO2 at bottling, approximately 30 mg/L, is required to suppress LAB activity in the presence of sorbic acid
• Warm storage temperatures accelerate LAB growth and sorbate metabolism; cool cellar conditions slow but may not prevent taint development over time

The simplest and most reliable prevention strategy is to avoid potassium sorbate in any wine that has undergone malolactic fermentation. Where sorbate is used, it should be accompanied by adequate free SO2 to suppress residual LAB. Sterile filtration to remove bacteria before bottling is the most direct physical approach. Lysozyme, approved by the OIV at up to 0.5 g/L, offers a targeted enzymatic alternative that kills gram-positive LAB without introducing sorbate's chemical risk. Producers committed to sorbate-free winemaking rely on combinations of sterile filtration, SO2 management, and selected LAB inoculants.

• Never add potassium sorbate to a wine that has undergone MLF; complete any desired MLF before sorbate is considered
• Lysozyme (up to 0.5 g/L per OIV; approved in EU, USA, and many New World regions) degrades LAB cell walls without the risk of geranium taint
• Sterile filtration through membranes of 0.45 microns or finer before bottling physically removes LAB, eliminating the biological precondition for the fault
• When sorbate is used, combine it with sufficient free SO2 at bottling; SO2 suppresses the LAB that sorbic acid alone cannot control

Potassium sorbate remains a legally permitted and cost-effective preservative in many regions, particularly for off-dry and sweet wines where residual sugar creates re-fermentation risk. The OIV maximum is 200 mg/L expressed as sorbic acid; the US limit is 300 mg/L. Its primary legitimate use is in stabilised sweet wines bottled without MLF, where the microbial conditions for geranium taint are least likely. Quality-focused producers in most regions now avoid sorbate in red wines and any wine that has undergone MLF, relying instead on filtration and SO2 management.

• OIV and EU maximum: 200 mg/L expressed as sorbic acid; US (TTB) maximum: 300 mg/L expressed as sorbic acid
• Appropriate use cases: residual-sugar white wines, dessert wines, and late-harvest styles where MLF has not occurred and yeast re-fermentation is the primary concern
• Sorbate should never be used in red wines, which are more likely to have undergone MLF and carry LAB populations
• Sorbate combined with adequate free SO2 reduces risk substantially compared to sorbate alone, since SO2 suppresses the LAB that sorbic acid cannot

Once geranium taint develops, it is practically irreversible. Ribéreau-Gayon et al. (2006) confirmed that the off-odour persists after activated carbon treatment, survives into spirits even after distillation, and remains perceptible after significant dilution. Laboratory confirmation uses gas chromatography-mass spectrometry (GC-MS) to quantify 2-ethoxyhexa-3,5-diene at sub-nanogram per litre levels. Sensory detection is straightforward for trained tasters given the compound's extraordinary potency. Affected wine cannot be corrected through any standard winemaking intervention.

• Sensory detection: a dominant, unpleasant aroma of crushed geranium leaves; easily recognised by trained palates at concentrations well above the 1 ng/L threshold
• Laboratory confirmation: GC-MS or GC-FID analysis quantifies 2-ethoxyhexa-3,5-diene at ppb or ppt levels in finished wine
• No corrective treatment is effective; activated carbon, fining, and aeration do not remove the compound; affected lots must be withdrawn or repurposed
• Prevention through sound microbial management, correct sorbate use, and adequate SO2 is the only viable approach; there is no remediation after the fact