Gordon Burns, founder of ETS Laboratories, speaks about cork testing Aug. 19 during the Wines & Vines Packaging Conference.Napa, Calif.

The problem with cork taint in wine has been well documented. By the late 1990s, an estimated 2% to 10% of wine bottles were believed to be spoiled by varying degrees of the taint, most often caused by 2,4,6-trichloroanisole (TCA). As a result, the cork industry saw some of its customers move to various alternative closures.

Clearly, something needed to be done. Peter Weber, executive director of the Cork Quality Council (CQC) in Forestville, Calif., approached Gordon Burns, co-founder and president of ETS Laboratories in St. Helena, Calif., in search of an objective way to test for TCA.

Burns described the research Aug. 19 during the second annual Wines & Vines Packaging Conference in Napa. The ETS research allowed the company to develop a better, faster test for TCA and helped the industry understand how the level of TCA in a cork translates into the amount of taint in wine. That research also has led to better testing of corks before they leave Portugal and has allowed the CQC to develop a protocol for its members that has greatly reduced the incidence of cork taint in U.S. wines.

A new way of testing

ETS started by hiring Eric Hervé, a chemist who was a recent Ph.D. graduate of the University of Bordeaux. The old ways of analyzing corks for TCA were laborious and not very good, Burns said, so they needed to develop a new method. “We turned Eric loose,” Burns said.

ETS had developed a technique using solid phase micro-extraction (SPME) before it was commercially available, followed by analysis with gas chromatography/mass spectroscopy (GC/MS). Hervé advanced that method, turning it into a tool that was much faster and could detect TCA below 1 part per trillion. “You’ve got to measure down to that range for it to be meaningful,” Burns said.

One key was understanding releasable TCA—the amount of TCA that is released when whole corks are soaked into wine. Hervé found that changes to the wine stop after about 24 hours. Not all TCA has migrated to the wine at that point, but an equilibrium is established, according to Burns’ presentation. “TCA likes to be in ethanol,” he said, “but it also likes to be in cork.” If you move the same cork to a fresh wine, TCA again migrates to the wine and reaches the same concentration as in the first wine, he added.

The next step was to find out if cork’s releasable TCA was at all related to TCA actually transferred into bottled wine. In 1998, Beringer Vineyards agreed to bottle white Zinfandel for ETS with about 400 corks that had been individually tested for releasable TCA. The findings: Releasable TCA was correlated to bottle TCA. After 14 months, bottle TCA was, on average, about 50% of the releasable TCA that had been found in corks. In other words, if a cork’s releasable TCA was 12 parts per trillion, the level in the wine was about 6 parts per trillion.

Burns and Weber took these findings to Portuguese cork producers in 2000. Weber said in a later interview with Wines & Vines that he wasn’t sure how the information would be received, but that the cork industry representatives asked them to continue the research. “I was surprised,” Weber said.

ETS eventually got a grant from the American Vineyard Foundation to do a larger study in 2004-06. Samplings were done on 10 cork bales with varying levels of releasable TCA, with the goal of determining whether group cork soaks were a good tool for quality control. Bottle TCA was determined after nine months and 20 months. A total of 2,700 SPME/GC/MS tests were done. Results confirmed the main earlier finding that releasable TCA was a good predictor of TCA-impacted bottles. Notably, the average transfer rate was found to be lower than in the earlier, smaller study. In fact, there seemed to be a threshold effect: Releasable TCA needed to be at least 4 parts per trillion in the cork for the level to reach 0.5 ppt in the bottled wine, which is detectable by GC/MS but not considered to have a sensory impact.

How the results are used

Meanwhile, based on the ETS research, the Cork Quality Council developed a testing protocol for its seven member companies, which include the major suppliers of natural cork. All incoming corks are subject to testing, and the sampling is based on an individual cork lot—defined as a single shipment from the same producer, same cork type, same visual grade and subjected to the same treatment. The ISO-2589-1 protocol is used to determine how many bales to test in each lot. Fifty corks from each bale are tested using group soaks.

Burns showed examples of results from individual bales. The Cork Quality Council currently considers any releasable TCA test exceeding 1.5 parts per trillion as a “flag.” Currently, nearly 90% of incoming lots of corks pass in the first round of tests. If the lot fails on the first testing round, a resampling that’s even more rigorous is triggered. Slightly more than half of the resampled lots ultimately pass. Cork lots that are rejected are returned to the manufacturer. Cork Quality Council members are subject to an annual audit.

Bar graphs showing releasable TCA levels in incoming corks over time, based on Cork Quality Council analyses, indicated that levels have fallen dramatically. In late 2001, the average was 4 parts per trillion; for the past several years, the average has been less than 1 ppt. Another graph showed that in 2002, barely 30% of incoming corks had their releasable TCA below 1 part per trillion, this proportion has increased to about 90% in 2014.

“Does this mean that the TCA problem with corks is completely gone? No,” Burns said at the conference. But he added that there has been a dramatic decrease in bad corks affecting consumers, as well as a big drop in the number of bad corks reaching the United States.

“I find it rewarding that (the research) has made as much of a difference as it has,” Burns said.