Cold stability,” as defined by the individual winery or winemaker, is achieved through cellar treatments. These treatments utilize aspects of the chemistry and physics discussed in Part I (Winter 2013 PWV Journal) to manipulate the formation of potassium bitartrate crystals.

Calcium tartrate crystal formation may also be affected by these techniques, but this salt requires more time to precipitate and is less affected by cold temperature methods. Stability with respect to both the potassium and the calcium salts is necessary to achieve crystal-free wines.

Several types of laboratory testing methods are used to evaluate the efficacy of the cellar treatments. Each cold stability laboratory test is empirical; every test is a predictive index of stability, not actual measures of stability. This limitation does not reduce their utility.

The various cellar treatments and laboratory tests use one or more of five aspects of impacting crystallization discussed in Part I. These are summarized simply as: 1) ionic concentration and complexing, 2) crystal formation and melting, 3) optimization of concentration and time, 4) diffusion, and 5) fouling.

By providing categorization of the various cellar and laboratory methods, and by discussing the issues involved in each method, we hope to provide common language to make any discussions about issues of “cold stability” more precise and productive.

HIGHLIGHTS
 

---

 
Berg, H.W., and M. Akiyoshi. 1971 "The Utility of Potassium Bitartrate Concentration: Product Values in Wine Processing." Am. J. Enol. & Vitic. 22: 127 -- 134.

Boulton, R. 1983 "Technical Notes: The Conductivity Method for Evaluating the Potassium Bitartrate Stability of Wines -- Part II" in Enology Briefs (G. Cooke, Ed.), Cooperative Extension, University of California. April/May.

Boulton, R., et al. 1996 Principles and Practices of Winemaking. New York: Chapman and Hall.

Dharmadhikari, M. Methods for Tartrate Stabilization of Wine. Iowa State University Midwest Grape and Wine Industry Institute Outreach and Extension. www.extension.iastate.edu/wine/methods-tartrate-stabilization-wine.

Guadalupe, Z. and B. Ayestaran. 2008 "Effect of Commercial Mannoprotein Addition on Polysaccharide, Polyphenolic, and Color Composition in Red Wines." J. Agric. Food Chem. 56: 9022 -- 9029.

Lubbers, S. et al. 1993 "Effect Colloide- Protecteur D'Estraits de Parois de Levures sure la Stabilité Tartrique d'une Solution Hydro-Alcoolique Modèle." Journal International des Science de la Vigne et du Vin 27 (1): 13 -- 22.

Moutounet, M., D. Bouissou, and J.L Escudier. 2009 "Determining the Degree of Tartaric Instability (DTI) Principle and Applications." Presented at Technology Day of the Champagne Oenologist Union, Epernay, France, Oct. 15.

Ribereau-Gayon, P., et al. 2000 Handbook of Enology, Vol 2: The Chemistry of Wine Stabilization and Treatments. New York: John Wiley and Sons.

Zoeckelin, B. et al. 1995 Wine Analysis and Production. New York: Chapman and Hall.

Zoecklein, B. 1998 "A Review of Potassium BiTartrate Stabilization of Wines" (Publication 463-013) www.apps.fst.vt.edu/extension/enology/downloads/PotBitar.pdf.

Cellar practices for cold stabilizing wines
Each cellar treatment has different demands on staff, energy, equipment and time. Further, each method has different potentials for quality risk due to oxidation, changes in acid balance or other sensory factors. Allowed processes and treatments for wine are listed in the U.S. Federal Register 27 CFR 24.246 and 27 CFR 24.248 and should be consulted before using any treatment or process.

New materials and processes, if they meet the standard of “good industry practice” and follow the TTB approval process, may also be considered for limited industry use. For example, some mannoproteins and carboxymethylcelluloses have been through the initial process and wineries may use them if proper paperwork is filed. TTB rules are elaborated at their website and should be consulted.

Cellar treatments focus on reducing the concentration of potassium tartrate (and calcium) by precipitation or other means, or on increasing the macromolecular chemical complexing of these ions and increasing the crystal fouling (the inhibition of the growth of the crystal via stearic or ionic hindrance). The methods listed reflect the current range and observed trends of stabilization methods.

Cellar Treatment Category I:
Ionic species reduction by precipitation methods
These cellar treatments use the precipitation of the salt crystals to reduce concentration of potassium (and calcium) cations and the bitartrate anion. Efficacy is increased when the crystallization is optimized through appropriate inputs.

A. Passive cold holding — exposing wines to natural cold weather extremes from winter weather, with no inputs, generally followed by separation of wine from precipitated crystals.

This method relies solely on the effect of temperature to induce possible crystal nucleation and precipitation, resulting in a decrease in the concentration of the ionic species. The ability for the salts to nucleate will be dependent on the concentration being high enough or on the existence of other nucleation sites (such as solid particles, ice or rough surfaces). This passive method expresses no control over the time and temperature exposures.

B. Active cold holding with inputs — refrigeration at a set or chosen temperature, with or without ice crystal formation or addition of cream of tartar (powdered or crystallized potassium bitartrate, i.e. “seeding”), and with or without mixing, followed by separation of wine from precipitated or added crystals.

These cellar ionic reduction methods allow for overcoming the nucleation and crystal growth energy and kinetic limitations of passive cold hold treatments through seeding with powdered bitartrate and/or mixing.

The effectiveness of these input enhancements can vary greatly, depending on the size and number of seed particles introduced and the ability to provide mixing capable of overcoming diffusion limitations. Completely engineered systems that control these variables and also quickly remove stabilized wine from crystals before any melting can occur are commercially available. This includes both batch tartrate stabilization and continuous tartrate stabilization (CTS) processes.

C. Rapid contact seeding — high concentrations of cream of tartar at refrigerated temperatures (but warmer than above), either using tank batch/seed/mix/filter techniques similar to above, or using crystallized beds, including fluidized crystallizers, with separation of wine from precipitated or added crystals.

Among the concentration reduction methods, these methods maximize efficiency on the crystallization side (with overwhelming nucleation and crystal growth power) to reduce the need to artificially concentrate the apparent ion concentrations through temperature drops. This allows for effective ion reduction at relatively warmer temperatures compared to traditional methods using less crystallization power.

D. Crystallizer and filtration technologies — maximizing temperature and seeding effects.

These are batch or continuous methods similar in principle to “active cold holding with inputs” and “rapid contact seeding” methods which aim to maximize all aspects of the underlying chemistry, kinetics, and physics. These methods maximize the effective concentration of the ions through temperature, use large volume of seed with very high surface area, optimize mixing, minimize fouling and quickly separate the wine from crystals prior to melting. 

Cellar Treatment Category II:
Ionic species reduction by other methods (non-precipitation)
Separation and resin-based exchange technologies allow for the reduction of  the ionic concentrations without relying on the precipitation of the salts. It has been proposed that these processes might disturb the equilibrium between ionic species and complexing agents. As elaborated in Part I, the ionic species can be weakly associated to molecules such as proteins or tannins, or even to metal ions. Some processes may disturb these weak bonds, resulting in more active concentrations of ionic species until the equilibrium is re-established, if at all.

A. Electrodialysis — removal of potassium and bitartrate ions through separation methods.

This process uses charge and membrane technology to selectively remove potassium and bitartrate ions (with a relatively minor loss of related species). This method directly lowers the available ions for salt crystal development.

B. Ion exchange — cation exchange (sodium or hydrogen ion exchanges for potassium and other wine cations), anion exchange (organic anions or hydroxide for tartrate ions) or a combination of anion and cation exchange.

Ion exchange reduces the concentration of the potassium, calcium and/or tartrate ions. Cation or anion exchange may be receiving renewed interest due to recent commercial enhancements in the size of ion exchangers and in the cost of resins available.

Cellar Treatment Category III:
Ionic species complexing and crystal fouling methods
A relatively new option for stabilizing wines, these methods do not reduce the actual concentrations of the potassium (or calcium) and tartrates. They reduce their effective concentration through chemical complexing and increased crystal fouling. Because these colloidal complexes are fragile, many wine processes, such as chilling or warming, filtration, fining, mixing or oxidation could disturb these complexes.

A. Increased yeast lees contact time — maximizes natural (in situ) mannoprotein extraction/concentration.

Originally applied to sparkling wines, and then observed in wines with extended sur lies aging, the increased lees contact resulted in increased in ionic complexing and crystal fouling.

B. Addition of commercially available/approved colloidal materials — including mannoproteins, carboxymethylcelluloses or acacia gums (or meta-tartaric acid in countries that allow its use).

These macromolecules act on several levels. They may interact with the ionic species to effectively decrease their apparent concentration, or may impede the formation of the nuclei, or may impede integration of the ions onto the crystals through steric hindrance by association with the ions, or may “foul” the crystals.

Commercial application of complexing agents require strict adherence to manufacturer’s recommendations, as these are rarely “simple” additions and entail planning, proper sequences and attention to temperature, filtration status and other wine treatments for best effect.

Laboratory testing methods to confirm cold stability of wine
All tests for cold stability are a “predictive index” and should be approached with an understanding of what they are predicting and how the test relates to the actual wine/situation/expectation relationship. Some of these methods attempt to predict if a wine will not precipitate crystals under any condition; others try to forecast for specific circumstances; while others methods limit their predictive modeling to the actual conditions of the lab test.

The current range and assortment of laboratory testing methods and indexes can be categorized into chill methods, conductivity methods and solubility methods, with variations within each type.

Lab Testing Category I: Chill methods
These relatively simple tests use a drop in temperature over time as a predictor of a wine’s likelihood to precipitate crystals. The nature of these tests makes them a poor index of calcium instability due to the short time and lack of any kinetic enhancers.

A. Freeze test
The most basic version of this test is to freeze a wine sample, then thaw it and look for the presence of crystals. Common variables of this test include the pretreatment of the sample (filtration compared to none); addition of KHT powder or not; sample size; freezer temperature; length of time in freezer (hours to days); thawing regime; temperature of evaluation and method of evaluation (visual compared to turbidity).

Test time commonly ranges from several hours to several days, with “overnight” being a common descriptor. Freezer temperatures also vary; the sample must change state from liquid to a solid. As a rough rule of thumb, the freezing point in: ºC = -0.5 x % alcohol (v/v) (Table I). Ultra-clear wines may occasionally refuse to freeze, invalidating the test.

The freeze test effectively increases the concentration product (CP) value by raising the alcohol and concentrating the potassium ion (K+) and the tartrate ion (HTa-). It induces nuclei through the increase in the CP compared to the holding capacity and kinetic limitations of the wine, to overcome the energy barrier for nuclei formation and/or use of ice crystals or scratches in the container surface for heterogeneous nucleation sites.

Questions about the freeze test involve how to deal with crystal formation and melting, the optimization of concentration and time and the lack of control of diffusion. The extreme nature of the test—actually changing physical state—makes the predictive index hard to understand. False negatives and false positives are definite possibilities. Despite these concerns, it is a commonly used test due to its speed, low expense and ease.

B. Cold-hold test (with or without inputs)
This test involves chilling the wine for an extended time, then evaluating the wine for crystals. These tests differ from freeze tests (above) in that the wine does not solidify or become slushy, and the time stored cold is generally longer, from several days to several weeks (or several months in some cases). Many of the cold-hold challenge tests are listed at -4ºC.

However, in actual practice, the cold-hold temperature will become that of the winery’s laboratory refrigerator, which may or may not be properly documented.

Common variations of this test include pre-treatment of the sample (filtration or not); additions of KHT powder to overcome the lack of seed in a filtered wine; different sample sizes; a wide range of refrigerator temperature and fluctuation; a variable length of time in refrigerator (days to weeks or months); different warming regimes; different temperatures of evaluation; and various methods of evaluation (visual or turbidity).

The “cold-hold” test may most closely mimic, on a short time scale, the actual situation a wine must survive, if the treatment inputs are properly defined and managed. As a predictive index this imitation of potential actual condition has great appeal, but the shorter time frame can cause concern, as a falsely reassuring test could result.

Cold-hold tests were used to develop many of the predictive models for faster tests. Cold-hold, with and without seeding, and held for extended time periods (months) at various temperatures, was used as the method to obtain the concentration product and the degree of saturation models. Three days at -4ºC (24.8ºF) cold-hold was used in modeling the “DIT” test, while six days at -4ºC cold-hold was used in forming the “ISTC-50” tests.

Lab Testing Category II: Conductivity methods
This general category of test uses the conductivity method as an indirect measure of potassium ions; conductivity can follow the change in potassium concentration in the sample during treatment. Conductivity is used rather than primary potassium analysis because is it rapid, relatively inexpensive and can be monitored “real time” during the test.

Conductivity can be affected by fluctuations of hydrogen ion concentration (changes in pH) or uncontrolled changes in temperature. Use of potassium testing to confirm conductivity results is limited by the current precision of potassium measurements (although this is changing rapidly), but could certainly be used as a research or confirmation tool when conductivity may be affected by these factors.

Conductivity tests have had only limited success in determining the calcium tartrate stability primarily due to the lack of effective nucleation. As currently used, most of these tests make few claims on being effective calcium tartrate stability predictors.

A. “Mini Contact” Conductivity Test, Variation I: “UC Davis Conductivity Test”
This test measures the change in conductivity over time of a chilled, mixed and heavily super-saturated (using powdered KHT “spiked” into the sample) wines. Common operator-chosen variables include test temperature; quality, mesh size, and amount of seeding tartrate powder; time of running the test; and acceptable conductivity change (either in percentage or absolute values).

Equipment variables include the ability to hold the selected temperature during the test, mixing intensity and speed, and choice of meter and probe sensitivity and ruggedness. All of these variables may impact the test. The published method specifies to run the test until two or three consecutive measurements are the same (usually less than 30 minutes).

The perceived simplicity of the test has been the major source of mis-application. There is tremendous variation in the quality of the equipment and the ability to control or document the critical test parameters, resulting in poor method performance.

The scientific appeal of this test is that it removes the kinetic limitations and, by choosing the temperature appropriately near the lowest expected temperature the wine is to experience, models it to that situation.

Recent commercialization of the cooling/mixing/conductivity system by both French (the “Stabilab” available from Oenodia) and Italian (the “Checkstab” available through Alpine Scientific) equipment manufacturers have introduced more precision into temperature control, conductivity measurement and time recording as it applies to the application of this test method. These (and other) brands of instruments are capable of running various versions of conductivity testing. Most can run the “Davis” test with much better overall control, and may result in improved performance while still offering the advantages of the method.

B. “Mini Contact” Conductivity Test, Variation II: “Degree of Tartrate Instability” (DIT)
This test measures the change in conductivity over time of a chilled, mixed and heavily super-saturated (using powdered KHT) sample and calculates the change in conductivity to infinite time using a proprietary program. Common variables include the temperature of the test, although it is usually run at -4ºC.

This method aims to avoid application problems of the “UC Davis Conductivity Test” by selling the method with its own instrumentation and removing as many variables as possible.

This test is considered “proprietary” and is generally run on one brand of equipment (Stabilab brand), which also supplies its own pre-weighed capsules of standardized bitartrate powder. This test is used to allow electrodialysis companies to “dial in” the amount of treatment needed, and for that reason it was set up to be strictly controlled.

C. “Mini Contact” Conductivity Test, Variation III: “Index of Tartrate Stability” (ISTC50)
This is another “proprietary” conductivity test from Oenodia, specifically selected to confirm electrodialyzed wine stability. Again, the only variable is the temperature, although the recommended test is done at -4ºC.

This is a conductivity-type test, but it does not use super-saturation of KHT powder to promote precipitation. Rather, it uses precisely quantified added KHT levels, said to be just enough to facilitate production of nuclei in what should usually be a stable, electrodialyzed wine. The precision is needed because the exact impact of this added bitartrate on the conductivity is subtracted from further calculations. The amount of spiked powdered tartrate used (0.5 g/L) is the “50” in the “ISTC-50.”

There is also use for ISTC-75 mentioned in literature. The rationale for the complexity of this particular degree of spiked bitartrate addition was to design a predictive test that would correlate to a three-day cold-hold at -4ºC. The justification for this test is that electrodialyzed wines fail other conductivity test methods; one proposed explanation involves a pH shift that artificially drops the conductivity value.

D. Degree of Saturation (Tsat)
The Degree of Saturation is a predictive index formed by the evaluation of hundreds of wines to determine the relationship between the ability of the wine to dissolve/absorb additional bitartrate (a process not subject to the kinetic challenges of the opposite direction) and its tendency to form crystals in cold-hold tests.

This index is determined by measuring the conductivity of a wine twice. The sample is tested without added excess powdered tartrate crystals and then a second sample is tested with a spike of 4 g/L powered bitartrate. The samples are warmed, and as any added tartrate crystals dissolve, the conductivity increases more than in the unspiked sample.

The difference in conductivity between the two samples at 20ºC is used in the Wurdig or the Maujean equation to calculate the Degree of Saturation (in ºC). This value, “Tsat,” estimates the temperature at which the wine is capable of dissolving additional tartrate. Additional empirical relationships are used to estimate the temperature of instability (the temperature at which this wine would be predicted to precipitate crystals) using the degree of saturation.

Lab Testing Category III: Chemical solubility methods
A predictive index based on solubility chemistry. Most cellar methods rely on the reduction of concentration of the relevant ions; this method focuses on quantifying those components.

A. Concentration Product (CP)
The concentration product (CP) test measures the concentration of the tartaric acid and calculates the bitartrate concentration using the pH of the sample. The relevant cation (potassium or calcium) is also measured; alcohol is analyzed to calculate its effect on solubility and activity. These values are used to calculate the value for CP for potassium (or the equivalent for calcium):

CP = [K+] [HTa-] = [K+] [H2Ta] x (% HTa-)

Tables were generated by determining CP values on wines treated by extended cold-hold with and without seeding. These tables are used to compare the CP values of the test wine. Variables include wine type, color, style and desired temperature of stability. This test only measures stoichiometric values, but the CP index also accounts for extended cold-hold test wines, which indirectly would account for some other stability aspects.

The impact of colloidal complexing molecules is the primary difference between the CP and the chemical solubility product (SP). CP methods also attempt to predict calcium tartrate stability.

This test may not be appropriate for any wines treated with complexing or colloidal materials unless another set of experimentally determined CP values are generated for a wide range of wines and situations. Similarly, there is concern that the chemical composition of wines used in developing the tables in the early 1970s may not reflect the chemistry of modern wines, especially with respect to colloidal composition.

However, CP values can also be used as confirmation tests for wines that have been stabilized (perhaps, for example, using a mini-contact test) to verify that cellar scale equilibrium has reached the desired concentration, rather than using the CP value tables derived from other wines. This allows for the development of “in-house” predictive indexes. Analytical method uncertainty must be considered when developing models, as any errors, particularly in potassium or tartrate concentrations, will be amplified.

Summary
The chemical equilibrium and kinetics of potassium bitartrate or “cold stability” in wine is not simple, and at least five issues must be considered whenever evaluating cellar or laboratory methods to understand if all known limiting factors have been addressed. Even the most cold-stable wines may still throw precipitates, either due to changes in the nature of the complexing molecules over extended time (a common event in red wines) or from precipitation of calcium tartrate.

The various laboratory tests are empirically derived predictive indexes, and each has proponents who have used them for many years. Although from a theoretical standpoint all the tests have inconsistencies or application issues, they are in daily use in the wine industry.

Many winemakers find reassurance over many years of using a specific test for what “expected” results should be and have derived experiential parameters for pass/fail levels for their chosen predictive index. Challenges occur when personnel, winemaking techniques and analytical parameters change.

The recent introduction of colloidal additives and electrodialysis treatments have, in particular, demonstrated that an understanding of both cellar treatments and laboratory methods are required to appropriately apply these new technologies.

Patricia Howe has more than 30 years of winemaking production experience as winemaking technical director at Domaine Chandon, Mumm Napa and the late Allied Domecq/Beam Wine Estates. She has a BS in fermentation science and a MS in food science (sensory science emphasis), both from UC Davis, and also worked there as a teaching laboratory manager. Her past business experience includes being owner/president of ASCENT Laboratories LLC, an applied sensory and analytical wine laboratory in Napa, Calif., and working at ETS Laboratories. She is currently the owner/winemaker of Patricia Howe Wines, adjunct instructor for VESTA and review process manager for the American Vineyard Association and a lecturer in enology at the Cornell University Department of Food Science in Ithaca, N.Y.