There’s a big effort to find new, commercially exploitable deposits of hydrogen. Ongoing exploration stretches across more than 20 countries, from Australia to Spain and on to Canada.

But these investigations are reminiscent of 19th century petroleum exploration – whether in Pennsylvania or Baku, early oil drilling took place near seeps that had been known for centuries. No one had the slightest idea that, in the 20th century, enormous volumes of petroleum would be found in the Arabian peninsula, the North Sea and offshore Brazil. Those discoveries had to await a deeper understanding of how, where and why hydrocarbons accumulate.

Current natural hydrogen exploration is at a similar, early stage of development, as it’s based on the equivalent of following oil seeps.

For example, current exploration in Australia and Brazil has looked for hydrogen leaks at the Earth’s surface, in the form of ‘fairy circles’ – km-wide patches of disturbed vegetation which indicate hydrogen accumulating in the soil. Alternatively, previously-drilled oil or water wells were occasionally analysed for the presence of hydrogen. This is how the discovery in Mali was made and is also the basis for current searches in the Pyrenees.

So, for now, we’re looking in places where hydrogen has already been detected. This makes sense, but uncovering the hydrogen-equivalent of Saudi Arabia may require a more science-based approach.

Fortunately, the techniques already used for hydrocarbon exploration tell us what a scientific exploration strategy would look like. Whether we’re after oil, natural gas or hydrogen, we need a source rock to generate it, a porous rock to hold it and an impermeable cap rock to trap it. A successful accumulation also needs a migration route connecting source to trap and, in addition, everything has to be in place at just the right time. There’s no point in having an otherwise fabulous trap that only formed a million years after the oil, gas or hydrogen passed by.

What do we need to find natural hydrogen?
Source rock, reservoir rock, cap rock, a migration pathway and good timing are, therefore, the necessary components for successful hydrocarbon exploration. Hydrogen exploration only really differs when we consider the source.

For oil and gas, our sources are ‘dirty rocks’ or, in more technical language, organic-rich sediments. Cook these at high temperatures and pressures for a few million years and crude oil or natural gas oozes out. What’s the equivalent for hydrogen?

There are several possibilities but the largest potential source, by far, is the Earth’s mantle – the 85% of Earth’s volume that lies between the core and the crust. Mantle rocks generate hydrogen when they’re warm and wet. Warm, here, means in the range 200–600°C. Wet means lots of free water (hydrated minerals don’t count as the water is chemically locked up within them). With this combination of temperature and moisture, water oxidises the reduced iron that mantle rocks abundantly contain, and hydrogen is released.

However, most of the mantle is hot and dry rather than warm and wet. Useful source rocks are, therefore, only found in places where geological processes cool down mantle rocks and, simultaneously, expose them to water. Fortunately, there are many plate-tectonic settings where this happens.

The classic example is the mid-ocean ridge system – the 65,000 km long mountain chain that runs along the centre of the world’s oceans. This ridge is where new crust is formed as brand-new tectonic plates are pulled apart by powerful convection currents deep within our planet. The resulting thin, fractured crust allows mantle to approach the surface and allows seawater to percolate deep beneath the seafloor. Warm, wet mantle is, therefore, found along much of the ridge’s length and this liberates vast quantities of hydrogen.

But this all happens in deep water and a long way from civilization – they’re called mid-ocean ridges for a reason! These ridges are unlikely, therefore, to form the basis of a new, clean, energy industry. Ideally, we want a hydrogen source on land and close to industrial centres, so that we can get it out cheaply and don’t have to transport it too far.

No one had the slightest idea that, in the 20th century, enormous volumes of petroleum would be found in the Arabian peninsula, the North Sea and offshore Brazil.... Current natural hydrogen exploration is at a similar, early stage of development, as it’s based on the equivalent of following oil seeps.

How do we find natural hydrogen?
There are other ways to moisten and cool the mantle. These generally involve a more complex sequence of geological events than occurs at mid-ocean ridges and, to quantify their potential properly, we need the hydrogen-exploration equivalent of the ‘basin modelling’ software widely used in hydrocarbon exploration. These are computer models that simulate the formation and deformation of sediments and the associated generation, migration and trapping of fluids. My own Royal Holloway research centre is currently developing basin models of hydrogen energy systems, although this work remains at an early stage.

To focus our minds, we’ve been thinking about the Pyrenees mountains, where natural hydrogen has been found on both the French and Spanish sides of the range. These accumulations may be independently sourced and entirely unrelated, but it makes sense to begin by assuming a common source somewhere under the central Pyrenees. There are at least a couple of candidates.

Firstly, as tectonic plates collided to form the Pyrenees over the last 40 million years, mantle rocks were pushed upwards by the resulting thrust faults. Thousands of cubic kilometres of thrusted mantle can be seen today on seismic images at depths of only around 10 km (where it should have just the right temperature). And, as this is directly under the highly faulted central Pyrenees, it is likely that rainwater has trickled down to meet it. The hydrogen potentially generated by this much mantle would have the same energy content as a supergiant oil field. Of course, hydrogen generation won’t have been 100% effective and not all of the generated hydrogen will have found its way to traps. But even a small fraction of this total would still be commercially attractive.

There is another, even more exciting, possibility. As well as generating thrust faults which pushed rocks upwards to form mountains, the Pyreneean collision also subducted the lower part of the Spanish crust under the French mantle. Think of the Spanish crust as meeting the tectonic equivalent of a fork in a road, so that the top half went up and the bottom half went down. And crustal rocks contain a lot of water. The subducting Iberian crust will have warmed as it was pushed deep into the Earth, and released its water which then buoyantly rose into the overlying Gallic mantle.

In most such settings, the result is ‘arc volcanism’ (as seen in the Andes or in Japan) because adding water to hot mantle makes it melt. Mysteriously, in the Pyrenees (and also in the Alps) there has been no such volcanism. Our models suggest that water release was unusually late in the Pyrenees, and that the mantle wedge was cooler than normal by the time water reached it. Could it be that, under these circumstances, water oxidises iron and generates hydrogen rather than producing magma? We don’t yet know, but we’re working on it.

The potential quantities of hydrogen generated by this scenario are staggering. The mantle wedge could have produced 10–100bn tonnes of hydrogen, containing the energy equivalent of over 10 years of total, global oil consumption. Once again, not all the available hydrogen will have been liberated and not all of that will have been captured in subsurface traps. Even so, if this model turns out to be correct, the implication is that south-west Europe may be the Saudi Arabia of hydrogen I’ve been after.

Reality check
The current excitement over natural hydrogen may be misplaced. The rosy picture I’ve painted above is very much a best case. More sceptical views are often based on the observation that, if there’s really a lot of subsurface hydrogen around, we should have spotted it decades ago. There is something to this argument, but there are also good reasons why we may have missed it. For example, analyses of gases in oil wells are usually carried out using gas chromatography with hydrogen as the carrier gas. That’s like looking for moisture in salt by first dissolving it in a test tube of water!

Natural hydrogen exploration is a classic high-risk/high-returns problem. Hydrogen accumulations large enough to be commercially viable may not exist at all. Or they may be so rare that they don’t make much difference. Still, it’s worth the risk of looking anyway as the economic and environmental pay-off will be enormous if a significant number of rich hydrogen accumulations really do lie under our feet just waiting to be found.

• Further reading: ‘The white gold rush: 40 companies now in pursuit of natural hydrogen’. The buzz around natural hydrogen – dubbed white or gold hydrogen – is gaining global momentum as a ‘potential gamechanger’ in the hunt for cost-effective, low-carbon energy sources, according to the latest research from Rystad Energy.