Fuel cells, more like fool cells...
Hydrogen from water is essentially a myth, electrolysis is about 3% efficient so it would be extremely expensive to implement, the electricity cost makes it uneconomical and it would produce more CO2 than ordinary petrol.
Hydrogen from natural gas is the most common route. That is up to 80% efficient on a good day. Hydrogen is transported and stored at the fuelling station. Its then compressed (typically electrically) to the desired pressure level, as there is no common standard for tank pressure yet. The car typically uses 2,000 to 10,000 psi with higher pressures being more common.
This still emits CO2. About 9 to 12 tons CO2 released per ton of hydrogen produced...
http://www.aidic.it/cet/10/19/007.pdf The Hyundai FCEV SUV (
http://www.autocar.co.uk/car-review/hyundai/ix35/first-drives/hyundai-ix35-fuel-cell-first-drive-review#) drives 400 miles on 144 litres of hydrogen... or 360ml hydrogen per mile... how much does this weigh? I can't find any sources on this from a quick search but I've heard talk of kg's of hydrogen being used and pricing in kg too.
360ml = 1 mile go juice, Tesla Model S = 180gCO2/mile from UK power, if production of 360ml produces more than 180g CO2 (20g hydrogen required to produce that) it's not looking good for FCEV...
I don't know the figures so can't comment on if it works out.
Carbon capture could be implemented but it remains to be seen if that is practical.
Lets also not ignore the fact that fuel cells are still incredibly expensive -- around 85% the price of an FCEV... and there's that tank of highly compressed hydrogen sitting in your garage.
Now petrol burns. It can explode but it usually burns, reasonably fast but not an explosion unless well contained and with the right mix of fuel and air, or as a vapour. Lithium ion batteries melt and burn slowly if something goes really wrong, the fire in the Teslas was 99% plastic, glycol coolant and metals burning, the vent ports are at the front of the car and the resulting hot gases/flames will start ancillary fires. Hydrogen? What if there's a leak? Hydrogen doesn't burn. Hydrogen explodes... look how quickly Hindenburg went up... now imagine a FCEV with a leak that gets a spark, and there's plenty of those around.
One spark, one ground connection arc, and one leak somewhere...
Now I'm sure you can make a tank resistant to crashes and impacts. But ultimately you're stuck with physical limitations. Hydrogen atoms are the smallest out there. Hardly anything can 100% contain a hydrogen atom. The leak needn't be visible to the eye. A tiny manufacturing defect -- a bubble of air, maybe, trapped in the casing of the tank -- and you have a leak. That leak could be a fire risk.
The other question is, yes, the tank is crash resistant, but HOW crash resistant? Although you're probably not going to hit dangerou speed in a car that does 0-60 in 10 seconds any time soon, when the cars improve and we have fast cars using this tech, what happens at an 80mph impact? I don't expect anyone in the impacted car to survive, no car can guarantee that -- but what could happen if that tank does rupture? Explosion? Fire? Would that cause problems for other vehicles?