Via the Economist (probably paywalled; Karlsruhe Institute of Technology PR will do as well): the idea is that instead of steam methane forming to produce hydrogen and CO2, you crack methane directly to hydrogen and carbon. The advantage is no CO2 to dispose of; instead you have carbon, which presumably has uses.
From their PR:
experimental reactor that could demonstrate the potential of methane cracking and overcome previous obstacles [carbon clogging and low conversion rates]. The starting point is a novel reactor design, as proposed by Carlo Rubbia and based on liquid metal technology. Fine methane bubbles are injected at the bottom of a column filled with molten tin. The cracking reaction happens when these bubbles rise to the surface of the liquid metal. Carbon separates on the surface of the bubbles and is deposited as a powder at the top end of the reactor when they disintegrate. This idea was put to the test during a series of experimental campaigns that ran from late 2012 to the spring of 2015 in KIT’s KALLA (KArlsruhe Liquid Metal LAboratory). Researchers were able to evaluate different parameters and options, such as temperature, construction materials and residence time. The final design is a 1.2-metre-high device made of a combination of quartz and stainless steel, which uses both pure tin and a packed bed structure consisting of pieces of quartz. “In the most recent experiments in April 2015, our reactor operated without interruptions for two weeks, producing hydrogen with a 78% conversion rate at temperatures of 1200°C. In particular the continuous operation is a decisive component of the kind of reliability that would be needed for an industrial-scale reactor” said Professor Thomas Wetzel, head of the KALLA laboratory at KIT. The innovative reactor is resistant to corrosion, and clogging is avoided because the microgranular carbon powder produced can be easily separated…
So, all very jolly, and perhaps it can be made to work. Clearly by not burning the carbon you’re losing part of the energy of the methane; and you have to supply heat for the cracking; though it did seem to me that might be naturally combined with a solar plant. Funnily enough, Rubbia has worked on that, too.