French scientists have devised a method of analysing valuable old bottles of wine to test if they are fake or genuine.
Using a particle accelerator, more associated with unlocking the secrets of the Universe than those of red wine, physicists at the Centre for Scientific Research in Bordeaux have teamed up with some of the great wine producing chateaux of the region and a wine merchant specialising in rare wines.
"We wanted to find some objective scientific method for analysing the wine and authenticating it in cases where it's not entirely clear if the bottle is genuine," says Stephen Williams of The Antique Wine Company, which bankrolled the VinCert project to the tune of £100,000.
The secondary trade in fine wine, with vintages dating back many decades or even centuries, now exceeds £1bn a year.
Mr Williams got the idea after purchasing a case of old wine from a house in the south of France, including a bottle of 1900 Chateau Margaux, worth in the region of £10,000.
The record auction price for a single bottle of wine is £96,000 ($156,000) for a 1787 Chateau Lafite, which was reputedly once owned by America's third President and author of the Declaration of Independence, Thomas Jefferson.
"We wanted to establish a scientific mechanism to verify the wine that we buy in cases where we can't satisfy ourselves that it's absolutely genuine," explains Mr Williams.
He had heard about a technique carried out by scientists at Arcane, the transfer technology arm of the Nuclear Research Centre at Gradignan in Bordeaux, which was using an ion beam from a particle accelerator to test the authenticity of fine art.
"We wondered if this could be developed further, and saw an application for wine," he said.
The analysis works by measuring the X-ray radiation emitted when the glass bottle is placed in the path of an ion particle beam.
"The detector allows us to analyse the chemical composition of the glass," explains researcher Herve Guegan.
"From that, we can detect the age of the bottle and also where the glass was made."
As glass production methods evolved over time, the chemical composition of wine bottles changed.
Trace elements produce a characteristic signature - or fingerprint - for each bottle, as some elements were only used in glass making for a few years before being substituted for something else.
For example, manganese was used to produce the green colour of wine bottles only between about 1920 and 1957, as it came from Moroccan mines at the time when Morocco was under French rule.
Then bottle manufacturers switched to chromium after Morocco gained independence.
So it is possible to compare the chemical fingerprint of a suspect bottle with that of a genuine sample to see if there are any substantial differences that would suggest that the bottle is fake.
The composition also gives clues to the specific production methods and furnaces used in different parts of France.
So far, the project has been able to build up a database of the chemical components of glass bottles from the last 200 years, using around 150 authentic bottles of fine wine donated by the various chateaux.
But it does not take a master forger to work out that you could get around this technique by simply taking authentic empty bottles and filling them with plonk.
So to get around this, a second test on the wine itself was devised by Philippe Hubert at the Centre for Nuclear Studies.
It uses a gamma ray detector to study the levels of radioactive particles in the wine, in this case caesium-137, that have been present in the atmosphere since the era of atomic weapons testing began after World War II.
"The main advantage of this technique is we don't need to open the bottle to do these kinds of measurements," Professor Hubert relates.
"We just have to put the bottle close to or on top of the detector."
Using bottles donated from the chateaux, Philippe Hubert has built up a record of caesium-137 levels in wine across the second half of the 20th Century.
"In the wine," he says, "is the story of the atomic age."
The measurements show that caesium levels rise dramatically from 1951, reaching a peak at around the time of the Cuban Missile Crisis in 1962, and then dropping dramatically, reflecting the atmospheric test ban treaty agreed by Presidents Kennedy and Khrushchev in 1963.
The next spike in the data comes in 1986, caused by fall-out from the nuclear accident at Chernobyl.
"If you have an old wine, say from 1860, for example; if you see some caesium in such a bottle, then immediately you can tell that this bottle is a fake one."
Both the gamma ray and the ion beam analyses should make forging antique wines much harder, particularly as the database expands to cover several hundred wines and include other regions such as Burgundy.
But some critics have pointed to a potential conflict of interest, given the close involvement of the chateaux and the financial backing of a wine dealer with a vested interest in authentication.
"Our objective here," explains Stephen Williams, "is to develop the technology and then hand it over to an independent company that will offer a service of authentication to other merchants and private individuals around the world."
Today, the best chateaux in Bordeaux, the so-called Premier Gran Cru Classé, take forgery much more seriously than in the past.
At Chateau Margaux, assistant winery manager Philippe Bascaules casts his eye over their unique and priceless collection of vintage wine, a reference cellar of authentic first growths from 1848 to the present day.
"Thirty years ago, I think we didn't care about the authenticity of the bottle," he says.
"It was unthinkable that someone could make a copy because the cost of the bottle was not so high. Now, when one bottle is more than 1,000 euros then people want to make one bottle of Chateau Margaux."
Margaux, like many prestige winemakers around the world, are taking fraud much more seriously now.
"We want it to be more difficult to copy a bottle of Chateau Margaux than other wines," says Mr Bascaules.
"So first of all the bottle has the logo of Margaux, but now we also put the vintage of the wine into the glass bottle. Since 2005, we also put a tag inside the bottle cap, and with a special reader we can recognise our cap."
Other technologies developed in the last few years include watermarks and holograms on labels and the use of special inks.
Hardy's, the Australian wine producer, has been encoding the caps of its niche wines with samples of DNA from 100-year old vines.