Adjusting Acidity With Membranes
The cool, rainy vintages of 2010 and 2011 have resulted in a comic assortment of pH and TA conditions that have sent winemakers back to their schoolbooks to relearn the basics. When pHs wander into the 4s during a cool, long season, sometimes the culprit is high malic (under-ripeness), sometimes high potassium (K+) (over-ripeness). Treatments differ entirely. In 2011, it wasn’t unheard of to get both conditions in the same must.
In this we have joined the ranks of our Eastern winemaker brethren now laboring for fully half the wineries in North America, and for whom expertise in this field is their chief employment qualification in the chilly northern climes of the Midwest.
The front office may sing love songs of non-intervention, but every winemaker knows that a dry white wine with 12 g/L (grams/liter) or a brown, tired, dried-up Pinot will fight an uphill battle to please the most luddite consumer. Sure, they want natural, and we want to give it to them. But above all, they want tasty wines and won’t settle for less.
Fortunately, today’s winemaker is blessed with a broad and rapidly expanding array of choices for bringing wines into acid balance, and a number of membrane techniques have recently been added to the dizzying array of tools. Choosing among them requires a firm grasp of acid-base chemistry.
The basics about acids
An acid is just a dissolved substance that can slough off a hydrogen ion (H+), really just a proton. Because the acids in wine are “weak,” some portion of the acidic hydrogen ions (protons) remains bound to them.
To figure when to pull the trigger on harvesting a block, winemakers look at both TA and pH. A glance at this season’s must chemistry will convince the most unschooled winemaker that pH and TA are not very closely related.
Both measure acidity (high TAs and low pHs denote lots of acidity), but TA is the sour taste, while pH is the amount of free, dissociated protons controlling the wine’s chemistry and microbiology. TA is like the cops on the payroll, while pH is like the cops on the beat fighting crime. TA is higher in tart wines, but low pHs mean high free acidity.
In normal maturation, malic acid is burned inside the berry, furnishing energy that grapes use to concentrate sugar into the fruit and lower TA. While TA is dropping, pH is rising from around 3.0 to as much as 4.0.
The strength of an acid is determined by the particular acid’s pKa, (i.e. the pH at which half of the acid is ionized). Lactic acid, for example, has a pKa of 3.8. Above pH 3.8, it’s mostly ionized, and below 3.8, it’s mostly undissociated.
Some acids have multiple pKas. Grape juice has two diprotic acids: the stronger tartaric acid (symbolized as H2Ta) with pKas of 3.0 and 4.2, and the weaker malic acid (H2Ma) with pKas of 3.5 and 5.0. Big difference. At wine pH (3.0-4.0), tartaric acid is always a lot more ionized than malic acid. Malic acid is like a really good donut shop where the proton cops like to hang out instead of patrolling the wine. In geek-speak, we say that the solution is heavily buffered.
A lovely characteristic of tartaric acid is that bitartrate precipitates with potassium to form crystals that reduce TA—very handy. This effect is maximized at pH 3.6, the peak of the bitartrate curve. This turns out to be a big deal, the natural great divide for winemaking.
Here’s something weird: Above 3.6, KHTa (potassium bitartrate) precipitation lowers TA and raises pH, just as you’d expect. However, below 3.6, TA is lowered but pH is also lowered. The acid goes down but it also goes up, resulting in softer taste, but with more stability and freshness. Not too shabby.
Because of this effect, it often works to de-acidify low-pH wines by adding potassium carbonate (K2CO3). This is always Plan A. Here’s how it works. First, as the compound dissolves, it ionizes into potassium cations and carbonate anions:
K2CO3 => 2K+ + CO32-
Next, the carbonate neutralizes some protons, benignly turning that nasty acid taste into carbon dioxide bubbles and water:
CO32- + 2H+ =>
H2O (water) + CO2 (bubbles)
Since it removes free protons, this reaction raises the pH. But the K+ ions will enhance bitartrate precipitation. As long as this happens below pH 3.6, this precipitation will lower both TA and pH, moving us back in the direction of the original low pH. A wine with a TA of 10 g/L and a pH of 3.1 can emerge with a TA of 7.5 g/L and a pH of 3.3 at negligible cost. Hot stuff.
Bad acid trips
If we have a lot of tartness in a wine, we might expect a nice low pH. Lots of critics and master sommeliers think that tart wines age longer. ’Tain’t necessarily so.
The wrong mix of acids can give you very tart wines with very poor shelf life. Grapes can get out of sync, so you get really high pHs when you still have high TAs. A high TA means the juice has lots of protons and a sour taste, but the high pH means they aren’t free and available. This can only happen if your protons are tied up somewhere. You have a lot of cops off duty, or in donut shops.
In a typical case, a juice may have a TA of 10 g/L and a pH of 3.9. Normally in California, the culprit is a high amount of potassium and tartrate. Tartrate is not a very good donut shop, but it will do the job if there’s enough around.
In such a wine, we have lots of K+, lots of tartrate, so we might expect a big precipitation. But high tartrate will not readily form KHTa if the pH is too far above 3.6.
The other way juice can have high pH and high TA is high malic acid. This happens all too often in Europe and North America, but it is rare in California except during chilly years like 2010 and 2011, when it’s anybody’s guess. Since malic acid is not easily removed, the first step is to determine whether it’s our problem at all.
Lab analysis for potassium and malate is expensive and time consuming, but there’s a simpler way. To test for this condition onsite, dissolve some tartaric acid in a small amount of warm water and simply acidify a sample of the juice to exactly pH 3.6.
Now freeze the sample overnight (allowing for ice expansion), then thaw it out in the morning. Hopefully you’ll see lots of white crystalline powder in the bottle. Either centrifuge, filter or settle out in the fridge, then run a TA. If the problem was high potassium, the resulting juice will have a big drop in TA to maybe around 8.5 g/L and a pH still at 3.6. If it works, go and do likewise to the big tank. If not, read on.
Getting the bugs out
Before we get on to high-tech membrane solutions, let’s discuss biological solutions. Organisms that eat acid have great appeal to our inner cheapskate. I will only speak generally here, because yearly advances in our knowledge promise to invalidate any specific information I might offer in this area.
Historically, biological de-acidification has been fraught with hidden costs and dangers. Thanks to the beloved and recently deceased Ralph Kunkee’s work at the University of California, Davis, in the 1970s, malolactic fermentation is a big success story, and today few winemakers are daunted by the prospect of pushing a wine through ML. But malolactic has huge impacts on style and is thus tricky to use to reduce acidity without harming fruit aromas. Most other biological solutions also create byproducts that alter style, so biological de-acidification methods must be evaluated with extreme care.
All saccharomyces cerevisiae yeasts consume some malic acid (generally around 10%) during primary fermentation without undesirable flavor production. Recently, strains like Lallemand’s 71B have received favorable marks in reducing as much as one-third of malic acid.
Pass the double salt
Were it not for the fuss and bother it entails, double salt de-acidification would be the standard treatment to reduce malic acid. It takes advantage of the precipitation of calcium malate that occurs at high pH. A portion of the juice (usually 20%–30%) is drawn off and treated with an excess of calcium carbonate (CaCO3). The carbonate reacts with 100% of available protons, both free and bound, completely neutralizing the juice to pH 7.0, while the TA drops to zero. Under these conditions, calcium precipitates both tartrate (CaTa) and malate (CaMa) in proportion to what is present, as well as its namesake double salt (Ca2TaMa).
When recombined into the main lot, a wine with a TA of 10 g/L will be reduced to 7.0, with 30% of its buffer capacity removed. The wine can then be re-acidified with tartaric if needed, restoring acid balance. The process does not create calcium instability because the final wine has only 30% calcium saturation.
Sounds good. The only trouble is, before recombining the treated portion, it is essential to filter it to remove all crystals and excess CaCO3, to say nothing of pulp solids—a slow, messy proposition. Cross-flow clarification to the rescue. The new tangential-flow filters making appearances all over the country to replace DE filtration seem tailor-made for double-salt filtration. Time to start sucking up to your neighbor who has one.
Double salt must be done prior to malolactic, and preferably at the juice stage, due to the hazards of taking a wine to such a high pH even for a short time.
Choose your weapon
Cross-flow clarification is emblematic of a dizzying array of new membrane technologies sweeping the wine world. Reverse osmosis (RO), an increasingly popular and available tool for removing rainwater from juice as well as for adjusting wine alcohol and VA content, has interesting prospects for de-acidification.
Reverse osmosis membranes used in the wine industry are made of the same materials employed in conventional sterile filters, but with pore sizes 10,000 times tighter. While sterile filters attempt to remove only particulates, RO membranes retain all but the smallest dissolved compounds.
We rate RO filters according to the molecular weight (MW) of a compound that passes 50% of its constituents into the filtrate (permeate).
I pioneered the use of tight reverse osmosis for the removal of volatile acidity. With a membrane MW cut-off of 80 daltons, only acetic acid passes into the permeate, to be trapped by a resin prior to recombining with the retentate.
The method takes advantage of the fact that ions are very large. The H2O molecule is a dipole attracted by the dozens to the charge on any ion, clinging like a gel layer that increases the ion’s functional molecular weight (FMW) by at least 500 daltons. The un-ionized acetic acid at 60 daltons will pass easily through an RO filter with an 80-dalton porosity, but its ionized acetate counterpart with a FMW of 600 daltons doesn’t pass through at all.
A precisely identical method may be used in de-acidification of excessively tart wines by employing looser RO membranes (near the 150-dalton legal limit) to pass malic acid at 134 MW. With a 150-dalton porosity, more flavor will be lost, but useful amounts of lactic and malic acids (pKas of 3.8 and 3.5, respectively) can be removed if the pHs are not too high.
The new generation of membranes focuses on membrane selectivity. Although the wine industry is tiny by global industrial standards, we are beginning to receive trickl e-down benefits from other industries, and off-the-rack technologies frequently appear and improve our options.
While RO membranes are impervious to ions, membranes have been developed that do just the opposite—ion-selective membranes that pass only the ions. Electrodialysis (patented by Eurodia and marketed in the United States as STARS), a method perfected some 20 years ago in France for economical and gentle cold stabilization, has been increasingly employed to great advantage for de-acidification.
In electrodialysis, wine is pumped between two membranes, a cation-permeable membrane that will only pass H+, K+ and Ca2+ and an anion-permeable membrane that passes tartrate and malate ions. A low-voltage DC current propels ions through these membranes—cations gravitating to the negative pole and anions attracted to the positive pole. In effect, KHTa is drawn into a brine that is discarded or sold.
In de-acidifying the high-TA wines of recent cool years, neutralization with potassium or calcium carbonate is limited by rising pH. If followed by electrodialysis, pH can be brought back down while simultaneously removing K+ and Ca2+ to prevent instability.
The beauty of this method is that unlike conventional cold stabilization, it protects the colloidal structure of the wine and saves a lot of energy. Electrodialysis can remove KHTa without the entrainment of colloids that accompanies crystal precipitation. Thus there is very little flavor stripping, and the method has been highly preferred to chilling by a trained sensory panel.
When used in concert with tartaric acid addition, electrodialysis can often give you virtually any desired pH and TA. Since it requires high clarity, electrodialysis runs on finished wine, even post-ML. Besides producing superior wines, electrodialysis can trim your energy bill by more than 25%. A system with an output of 400 gallons per hour costs $190,000. Because of its high capital cost, electrodialysis systems are usually accessed as a service except by large wineries.
Because it is less ionized due to its higher pKas, loose RO preferentially removes malic acid over tartaric. For the same reason electrodialysis, which only removes ions, is not very effective in removing malic acid. This is its Achilles heel—it is not the tool of choice for those overly crisp Midwestern whites.
An additional alternative selective membrane technology is currently being marketed by Mavrik Industries. CEO Bob Kreisher is frank about its proprietary nature. “We want winemakers to feel comfortable with what our process does, but we worked hard on developing a system that works well, and we don’t want to give away essential information to our competitors,” he said.
Mavrik was nonetheless quite open to my observing what their system does, and I got a firsthand look at the de-acidification of a 2010 Cabernet.
Mavrick’s technology is reportedly based on an acid-selective membrane that passes both molecular and ionized species. This is big news, greatly simplifying our lives because we no longer need to pay attention to ionization pH and malolactic status. Colorless, flavorless permeate from this membrane is passed through a weak anion exchange resin where malic, tartaric and lactic acids are retained. A cation-exchange column can also be employed to exchange potassium for H+, with the effect of re-acidfying the permeate prior to recombination.
In this way, buffer capacity is removed, and pH and TA can be adjusted more or less at will. Although I cannot claim to fully grasp the details of Mavrick’s proprietary magic, I can at least attest that the Cabernet came through it with flying colors and no discernible aroma loss. As Mavrick refines this technology, it may very well be the toy of choice in regions plagued by high malic acid.
Back in sunny California, extended hangtime often results in high-potassium wines that resist pH adjustment with tartaric. A new approach to acidification of high-pH wines and musts now being pioneered by Eurodia uses a bipolar membrane on the cation side to exchange potassium ions for protons, thus raising the TA. Unlike tartaric acid addition, this lowers pH without increasing buffer capacity. The membrane works much like a cation-exchange resin, removing potassium ions in trade for H+, but without the stripping of flavor elements that occurs during direct contact of wine and resin. Bipolar anion applications may also be on the horizon.
The bottom line
Any reporter loves a scoop. In preparing this article I was treated to a generous handful of new technologies that promise to transform American winemaking in an era of climate change, up-end modern winemaking precepts and render current teachings obsolete.
Yet there is no bottom line to report. To a man, technology developers waffled and temporized concerning release dates, performance, efficiencies and capital costs of their new darlings. The complex and peculiar machinations of TTB approval are an uncharted minefield through which wineries must walk with care and do their own homework.
We are smack in the middle of an era much like the 1960s, when a tsunami shift from Ports and Sherries to table wines left us in total ignorance. Today we are completely unprepared for the impact of de-acidification capabilities even their providers have yet to fathom. Smart postmodern winemakers throughout the country and their state-sponsored academic partners are well advised to place a high priority on understanding and evaluating the diverse menu of options soon to be thrown on our plates.
In all candor, this article provides no dependable Consumer Reports purchase guide to de-acidification. In its stead I have sought to lay the groundwork for such an evaluation by tracing a roadmap of the options we must immediately get smart about.
Clark Smith is winemaker for WineSmith and founder of the wine technology firm Vinovation. He lectures widely on an ancient yet innovative view of American winemaking.