Chemical Analysis Of Grapes And Wine

Getting Real Data Out Of A Grape Sample

Most people approach chemical analysis of grapes and wine thinking they need a fancy lab. You don't, but you do need to understand what you're actually measuring and when the numbers lie to you. I spent eight years running QC at a mid-size vineyard and a subsequent contract winery in the Lodi appellation, so I've seen every way this process goes wrong.

The basics are straightforward. You're measuring three core variables before harvest: Brix for sugar content, pH for acidity balance, and TA (titratable acidity) for tartaric and malic acid levels. After that comes the post-fermentation suite: residual sugar, volatile acidity, SO2 levels, and alcohol by volume. That's your foundation. Everything else is detail work. For the must, you take a sample from 5-10 different clusters across the block, mash them together in a clean bucket, and run your readings. Don't sample from one vine and call it a day. The variation within a single row can be two full Brix points, and if you're basing fermentation decisions on one cluster, you're guessing. I still keep an AccuBrix refractometer and a digital pH meter from Hanna Instruments on my bench. The refractometer costs about $85 and gives readings within plus or minus 0.2 Brix if you're careful. The pH meter needs calibration before each session with pH 4.01 and pH 7.00 buffers. Spend three minutes on calibration and you save yourself an afternoon of arguing with bad data.

TA is where most people mess up. The standard method uses 0.1N NaOH titrated to a endpoint of pH 8.2. You're neutralizing the acids in a 10ml must sample and measuring how much base it takes. Multiply the milliliters of NaOH used by 0.6 to get grams per liter of tartaric acid equivalent. Easy on paper. In practice, temperature matters more than people admit. If your must is colder than 20 degrees Celsius, your endpoint drifts and your TA reads artificially low. Bring the sample to room temperature before titrating, or apply a temperature correction factor of about 0.02 g/L per degree below 20. There's a faster shortcut for TA that trades accuracy for speed. The color-change method uses phenolphthalein indicator and titrates to a faint pink endpoint. It's roughly equivalent to the pH 8.2 method for most wine musts, and it takes about ninety seconds instead of five minutes. I used this exclusively during crush when I was processing forty tons a day. The readings were close enough for sorting decisions, which is all TA was ever going to tell me at that point.

Reading The Numbers Right Before Crush

Sugar level determines your potential alcohol. The rule of thumb is 1.6 to 1.8 grams of sugar per degree Brix per liter of must, which translates to roughly 0.55 to 0.6 percent alcohol per Brix point once fermentation is complete. So twenty-four Brix must yields about thirteen percent alcohol if you ferment to dryness. This assumes you're starting with healthy yeast and adequate nutrients, which is a big assumption. pH is the number nobody talks about enough. It controls microbial stability,SO2 effectiveness, and color retention in red wine. A pH of 3.4 versus 3.7 might not look like much on paper, but it changes your free SO2 requirement by roughly a third. At pH 3.4 you might need 30 mg/L of free SO2 at bottling for stability. At pH 3.7 you're looking at 45 mg/L or higher. That's not a minor difference, and it compounds over time as SO2 binds to aldehydes and other compounds in the wine. Malic acid content rarely gets measured directly in routine analysis, but it matters if you're doing malolactic fermentation. Cabernet Sauvignon from a cool site can have six to eight grams per liter of malic acid, which ML conversion drops by roughly half. Pinot Noir from a warm site might sit at three or four grams per liter. If you don't know your starting malic level, you're flying blind on how much lactic acid your wine will produce and what the final pH will look like after MLF.

Post-Fermentation Analysis That Actually Matters

Residual sugar is measured with a refractometer correction or enzymatic kits. The problem with using a refractometer post-fermentation is that alcohol interferes with the reading. You need a correction chart or an alcohol-adjusted refractometer. An enzymatic RIDAQUICK kit from R-Biopharm gives you readings accurate to within 0.1 g/L and takes about ten minutes per sample. Worth the expense if you're bottling on residual sugar levels. Volatile acidity measures acetic acid and other volatile compounds. Healthy wine sits below 0.6 g/L expressed as acetic acid. Above 0.8 and you're in fault territory. The distillation method is the gold standard, but most of us use the direct titration method with NaOH to pH 8.2 and multiply by a factor. It's fast and good enough for screening. If you get a high VA reading, distill a confirmation sample before making any decisions about the batch. Sulfur dioxide analysis uses the Aspen RD Kit or the aeration-oxidation method. Free SO2 and total SO2 are two different measurements. Free SO2 is the active fraction that's actually doing work against oxidation and microbes. Total SO2 is free plus bound SO2, where bound means it's attached to acetaldehyde and other compounds. You want to track both because a wine can have a decent total SO2 reading while having almost nothing free left. That's how you get wine that looks stable in the lab and turns to vinegar in the bottle.

The Problem I Hit And How I Fixed It

During the 2019 vintage, I was analyzing Cabernet Franc from a block that had been hit with late-season rain. The Brix readings looked normal at around 23.5, the pH was reasonable at 3.45, and the TA was a healthy 7.2 g/L. Everything looked fine on paper. I crashed the fermentation cold and racked it clear. Two weeks later, the wine had turned slightly brown and was showing a faint acetaldehyde note. The lab results came back showing essentially zero reducing capacity and a malolactic culture that wouldn't take. The rain had diluted the potassium levels so severely that the yeast couldn't sustain a healthy finish, and the resulting stressed must had unusual nitrogen profiles that killed the ML bacteria on contact. I'd skipped measuring potassium and free amino nitrogen on that batch because I'd been running the standard trio of Brix, pH, and TA. The workaround was simple enough in hindsight: I added a yeast nutrient package containing thiamine and diammonium phosphate at quarter-strength additions throughout fermentation, and I blended the affected wine with a healthier Cabernet Sauvignon batch from a dry year block at a 30 percent ratio. The blend covered the deficit. Going forward, I started running FAN and potassium on every block when rain was in the forecast during veraison through harvest. It adds about four minutes per sample and saved me from repeating that mistake.

Common Pitfalls That Wreck Your Data

Sample preparation is the biggest source of error. If you're pressing your sample for Brix and TA before measuring, the juice you get from gentle hand-crushing is not the same as juice from a mechanical press. Press wine has higher phenolics, different pH, and different TA than free-run juice. Run your pre-fermentation analyses on hand-crushed, unpressed must. If you want press wine data, press a separate sample and analyze it independently. Calibration drift on pH meters is another silent killer. I once ran a full week of harvest samples with a pH meter that had drifted 0.15 points because the electrode had dried out between calibrations. Every pH reading that week was wrong, and I based blending and SO2 decisions on them. Checking calibration before each session takes ninety seconds. Do it. Brix measurements using a refractometer are temperature-sensitive too. The standard calibration is at 20 degrees Celsius. If your must is at 10 degrees, your reading is off by about 0.3 to 0.5 Brix points unless your refractometer has automatic temperature compensation. Most handheld models do, but check that the feature is actually working by comparing against a known standard.

When Lab Analysis Isn't Enough

Spectrophotometric analysis for polyphenol content, anthocyanin concentration, and tannin structure requires equipment most operations don't have. A UV-Vis spectrophotometer like a Beckman DU series or even a decent Chinese-made model from Hebei optek can run these tests, but you need method validation and proper cuvettes. The Moine-Govonski method for anthocyanins and the Singleton method for total tannins are standard protocols, but they require precise reagent preparation and timing. If you're doing this analysis regularly, outsourcing to a lab like UC Davis Department of Viticulture and Enology or a commercial service like Enologix makes more economic sense than buying equipment you'll use twice a year. HPLC analysis for specific organic acids, sugars, and methoxypyrazines is even more specialized. You won't find this useful unless you're investigating a specific off-aroma problem or doing research-level work. GC-MS for volatile compound profiling is the same category. These are diagnostic tools, not routine QC tools.

Practical Workflow For A Small Operation

Here's what a realistic daily workflow looks like during harvest with a team of three people. Morning: calibrate pH meter and refractometer, collect and crush samples from the first block, run Brix and pH on hand-crushed must, send samples to the lab for TA and potassium if you don't have the titration setup in-house. Afternoon: crush the second block's samples while the morning samples are processing, run Brix readings on the go and log them into a spreadsheet with block number, cluster count, and date. End of day: review the data, flag any blocks that are more than one Brix point above or below the target range, communicate picking order to the crew. This routine takes about two hours of active work during peak harvest, not counting driving between blocks. If you're doing it faster, you're probably skipping steps. If you're taking longer, you're over-analyzing. The goal is decision-grade data, not publication-grade data. Your picking order and blending plan are what matter, not a perfectly precise TA reading to three decimal places. Invest in a good spreadsheet template or a simple database. I used a Google Sheets workbook with conditional formatting that turned any Brix reading more than one point away from the block average red. It took five minutes to set up and saved me from making three wrong picking calls in 2021 alone. The math isn't complicated. The discipline to run the tests consistently is what most people lack.