• Close but no cigar

    The machine that didn't work

    Dipayan Paul wanted to further explore a revolutionary technique to detect radiocarbon. Instead, he proved the technique doesn’t work at all and butted heads with the group that invented it in the first place.
    in short

    Isotope physicist Dipayan Paul went to America to further explore a revolutionary technique to measure C14.

    C14 measurements are used to date organic material up to 50,000 years in the past, but also to determine the amount of natural CO2 in the atmosphere as opposed to CO2 from fossil sources.

    The American method – called ICOGS – claimed to be easier and much cheaper than taking measurements with an Accelerator Mass Spectrometer.

    The universities of Groningen, Uppsala and Columbia banded together to develop and adapt the technique. However, they couldn’t replicate the results, not even with samples that contained ten to a billion times the natural concentration.

    Paul published his results and got support from the Uppsala researchers. But the Murnick Group still claims their technique works.

    Paul did two extra projects to still be awarded his PhD. ‘Failure is part of the game’, he says. ‘This has made me a better scientist.’

    full version

    Reading time: 7 minutes (1,300 words)

    They’re not exactly best friends anymore, isotope physicists Dipayan Paul and Harro Meijer from Groningen on one side and their colleagues from the Murnick Group at Rutgers University in New Jersey on the other. For years on end, the two pairs were in touch on a regular basis, but not anymore. Their only interaction nowadays is in the battle of cutting comments on an article that Paul and Meijer published seven months ago in Analytical Chemistry.

    Why? Well, in this article, Paul – who is receiving his PhD on Monday – states that a revolutionary new method to detect C14 that brought fame to the Murnick Group doesn’t work.

    ‘It’s very tricky’, says Paul. ‘I have worked there for eleven months and they were really supportive and enthusiastic. After that, I went back to Groningen and built our own device here to test it even further. But I couldn’t reproduce their results. A Swedish group from Uppsala couldn’t either and, to be honest, they couldn’t even reproduce them themselves. However, they never acknowledged that openly. So yes, by now, it’s gotten a bit nasty.’

    Easier and cheaper

    Murnick’s revolution-to-be was called ICOGS, or IntraCavity OptoGalvanic Spectroscopy, a method that was so promising that it seemed poised to replace the traditional Accelerator Mass Spectrometer, one of which is housed at Nijenborgh 4. The AMS is currently the most reliable and precise way to detect radiocarbon. It uses the tiny amount of radioactive C14 in organic matter to date it back to about 50,000 years in the past.

    However, the machine itself is expensive: two million euros. On top of that, it’s a very complex instrument. It requires a lot of training to operate it and the preparation of samples takes many steps. ‘For example: you have to combust the samples to CO2, which is then converted to powdery graphite and afterwards pressed onto an aluminum holder to do the measurements’, Paul explains.

    It’s no wonder that when the Murnick Group came up with this new method that was supposed to be just as accurate as AMS, but much easier and cheaper, Groningen – one of the leading C14 labs in the world – wanted to explore it further. The ICOGS device only cost 150,000 euros and is way smaller. And because it relies on the different ways carbon isotopes respond to infrared light, there’s no need to burn the CO2 to graphite.

    Not easy at all

    The universities of Groningen, Uppsala and Columbia banded together to further develop the technique and adapt it for their own purposes. The scientists at Rutgers University welcomed them with open arms. ‘They showed me everything’, Paul says.

    Younger-looking carbon

    One problem with detecting and measuring C14 in very old samples is contamination from modern carbon that contains much more C14, which makes the sample look younger. ‘It’s important to understand that process, because we perform high quality measurements. However, I found it wasn’t possible to completely prevent contamination. But you can take measures to minimise it.’ Those measurements have now been partially taken in the Groningen lab. Others will be able to be taken once a new AMS in the new science building becomes available.

    At Rutgers University, Paul started his own experiments. ‘The Americans showed the method to be effective for natural C14 concentrations’, Paul says. ‘So we initially tested it with CO2 with a contemporary amount of radiocarbon and a ‘dead’ sample, with almost no radiocarbon.’

    Both samples were representative of the research the Groningen group wanted to use it for. Yet Paul couldn’t detect the C14 signal in the samples at all. That’s when the trouble began.

    ‘We decided to make some modifications. We checked several parameters, like cell pressure and the wavelength of the laser. But none of the experiments showed any clear evidence of a C14 signal. That made one thing clear: the 2008 paper said the procedure was easy, but it wasn’t at all.’

    High levels of radiocarbon

    Paul concluded that the sensitivity of the device most likely wasn’t high enough, which meant the scientists at the RUG could probably never use it. But he still had to finish his PhD. He decided to continue his research by performing tests in Groningen, where he built his own machine to further examine the technique with heavily enriched samples. Those samples contained ten to one billion times the amount of C14 that naturally exists in the atmosphere.

    ‘Because of the high levels of radiocarbon involved, I couldn’t even bring the samples to our own lab’, he says. ‘I had to move the experiments two buildings away to avoid contamination in our lab here.’


    Carbon-14 measurements are also used to measure how much of the CO2 in the atmosphere is natural and how much is from fossil sources. In the latter case, it will contain less C14. It is also used as a tracer to better understand the global carbon cycle.

    In 2010, scientists at NOAA (National Oceanic and Atmospheric Administration) developed a very innovative atmospheric sampling method called AirCore.

    This involves bringing a special tube to high altitudes with a balloon and ‘dropping’ it. The tube collects air on the way down. However, the samples are extremely small. Paul wanted to know whether it was possible to extract and measure enough C14 from these samples. ‘We found it was possible and that AirCore is a very promising sampling method’, Paul says.

    By then, he was no longer trying to test the method but simply wanted to see how much C14 was needed to even get a signal at all. But much to his surprise, even at the highest concentration, he couldn’t get any results.

    ‘That was the last nail in the coffin’, Paul says. His supervisor, Harro Meijer, decided to stop the project. ‘The group from Sweden had already concluded it didn’t work with samples up to ten times the natural abundance.’

    A lot of fuss

    As Paul began working on two different projects which would enable him to still get his degree, he also wrote about his – negative – findings. The resulting article was published in August of last year. He’s happy he still got a publication in a very respectable journal out of all the work he did. But it certainly kicked up a lot of fuss that is still ongoing: it has gotten awkward, to say the least.

    The Murnick Group isn’t taking the criticism lying down. Instead, they wrote a comment on Paul’s paper referring to ‘inconsistencies’. ‘It is shown that their conclusions can be refuted with a more careful analysis of their data’, Murnick writes.

    ‘We commented on that comment, of course’, Paul says. It even came to the point that the editor asked the Uppsala group to comment on the paper. So they did – and they agreed with Groningen. Finally, all three comments were sent to three independent reviewers. ‘Last week, we heard from the editor-in-chief and received the reviewers’ comments. They agree with all our points and decided to publish our response and the comments online’, Paul says.

    He never wanted to cause any trouble. He just wanted to explore a great new technique and add to the expertise of the Groningen group. It just wasn’t meant to be. ‘The scientific community deserves to know the truth about the ICOGS.’

    It’s been a bumpy ride, Paul admits, but he has learned so much. ‘You know, when I applied for the PhD position here, they asked for a scientist who was motivated and independent. Back then I wasn’t sure what that meant, but now… It’s very different here than in – say – America. Here, they expect you to fail and learn from it. It’s part of the game. This has made me a better scientist.’