Imagine you're stuck in the Australian desert. Your car has broken down, and there's no sign of another person for miles.

A survival expert will tell you your first priority is finding water.

It's probably not a situation many of us will ever face, but for some people and in certain regions of the world, reliable access to clean drinking water can be a constant struggle, and droughts can be deadly.

So being able to effectively conjure water from thin air, as fanciful as that sounds, could dramatically change people's lives.

Earlier this year, a team of researchers from the University of California Berkeley reported they had created a device that can extract water from dry desert air, without using an external fuel source.

The drawback was that it used a zirconium-based powder called meta-organic framework 801 (MOF-801), which is expensive to produce and relatively inefficient.

But reporting in today's Science Advances, they say they have refined the design. The new device has an aluminium-based framework with more than double the water-capturing capacity, and can be made at around one-hundredth the cost of the zirconium-based equivalent.

How does it work?
The key to the water-capturing capacity of the design is the massive surface-area-to-mass ratio of the MOF powder, according to researcher Eugene Kapustin.

"It's very porous. It has an enormous amount of space inside. The surface area of one gram of this MOF can cover several soccer fields," Mr Kapustin said.

The device is opened at night when desert air is at its most saturated, allowing water to be absorbed by the MOF powder.

During the day when relative humidity can drop to 10 per cent or less, the insulated case is closed and an aerogel is placed over the top of the MOF powder, which absorbs heat from the sun.

When water vapour released from the heated powder comes in contact with a condenser which has been cooled overnight, it cools and condenses the vapour, allowing the liquid to be collected.

The zirconium-based MOF-801 design, which they field tested in the Arizona desert, was able to yield around 100 millilitres of water per kilogram of powder.

But lab testing has shown the newer MOF-303 to be far more efficient.

"We reproduced the desert conditions in our lab using the new meta organic framework 303," Mr Kapustin said.

"It was able to deliver up to 400 millilitres of water per kilogram of powder."

Australian researchers in line for global prize
The Berkeley team are not the only researchers working on this kind of technology.

An Australian team is among five shortlisted in a worldwide XPRIZE competition, which called for inventions which could help "alleviate the global water crisis".

The XPRIZE competition challenged entrants to "create a device that extracts a minimum of 2,000 litres of water per day from the atmosphere using 100 per cent renewable energy, at a cost of no more than 2 cents per litre".

Led by Professor Behdad Moghtaderi from the University of Newcastle, they will find out in August whether their design wins the $1.75 million first prize.

Similar to the Berkeley model, Professor Moghtaderi's system absorbs water at night into silica pouches. The pouches are then heated by solar energy during the day releasing water vapour, which is then cooled.

But unlike the Berkeley design, Professor Moghtaderi's unit uses solar energy to heat the air in the system, and to drive a fan and condenser, he said.

"The ambient air enters at about 25 degrees Celsius average, and exits our system at about 85 degrees," he said.

The hyper-saturated air is then pushed through a condenser where it is cooled and releases the water for collection.

Professor Moghtaderi's modular system is capable of producing about 20 litres of water per module per day, but he said the Berkeley design may be limited by the scale it can operate on.

The issue is that the system needs large volumes of dry air to come in contact with the MOF, but without a fan it is reliant on passive airflow.

"Let's says these guys are aiming to produce 20 litres of water per day. With their system, they'd need at least 100 kilograms of that [MOF powder], and they'd need about 800 cubic metres of air to get in contact with that material," he said.

"Where are you going to get the energy to bring moist air into the material?"

'Almost one third of the world is water stressed'
Although both models are yet to be tested at a commercial scale, the potential for a device that can produce water in arid conditions is vast.

The most obvious benefit would be in drought-prone regions or in areas where water quality is poor, according to Berkeley project leader Professor Omar Yaghi.

"Almost one-third of the world is water stressed," Professor Yaghi said, during a Berkeley interview.

And water stress is likely to become more widespread as climate change throws a greater level of unpredictably into weather patterns.

Cape Town in South Africa is still at risk of running out of water despite recent rains, and the Australian Department of Environment and Energy lists increasing drought as a key threat to Australia's agricultural future.

Mr Kapustin said the capacity to produce water at will could have both rural and urban benefits.

"Australia also experiences water scarcity. So it could be potentially useful for agriculture, when there is a need for fresh water," he said.

"In cities like Flint Michigan in the US, where there is water but it's always contaminated, these devices can produce drinking water."

On a smaller scale, it may be possible for portable versions to be carried when travelling in remote areas like central Australia.

A 25-year-old man died after his car broke down while working in the Simpson desert in 2012, and a Darwin tradesman claims to have been forced to drink his own urine after crashing his car near the Northern Territory–South Australia border last year.

Whether Professor Moghtaderi's team win the XPRIZE or not, he said they would continue with their research.

"I'm very happy that we made it this far and we are actually in discussion with a couple of companies in the water area who are interested in trialling our system and then perhaps commercialising it."