The really intriguing element to this idea is its ability to bridge from the existing gasoline economy to the fabled hydrogen economy. One of other big challenges for hydrogen-fuelled vehicles (besides storing enough fuel) is building out the necessary infrastructure and building up the base of fuel cell-powered cars in parallel. This technology could be a big help in accelerating adoption by enabling existing cars to be converted, a much cheaper and faster ramp-up that in turn will make fully electric fuel cell vehicles more attractive. A fascinating idea all round.

" [ ... ] a full tank of aluminum-gallium pellets, [ which ] amounts to about 350 pounds of aluminum

[ ... ] the added weight would be the equivalent of an extra passenger, albeit a pretty large extra passenger."

Not in USA it wouldn't be, that's about average for an American child these days isn't it?

;-)

So whats the actual efficiency of this once you include emptying the aluminium oxide from your tank at the petrol station, carting it back to a power generating facility, then hauling the new aluminium pellets back to the petrol station?

I'll admit if the electricity used is cheap and clean then this has a lot of benefits...

However only really iceland is going to have that advantage (a country where a lot fo aluminium ore is shipped to because of how expensive aluminium is to refine, given its huge power requirment) because iceland has such a surplus of geothermal power it because more economic to ship the ore then the ship the refined aluminium back...

would the same happen if this went into cars?

I still see a problem though... 175Kg of aluminium, plus whatever water capacity you need, vs the normal 60-80kg of the fuel tank?

Its not going to work too well in smaller cars, but then this did come from american research and they don't seem to do small cars over there

Is the violent poof like Ronnie Kray? Because we can do without that kind of thing being generated all that often.

Now we just need to find an energy-efficient way to produce all this aluminium. Hm, how about building some light oil reactors? After all, we won't need that gasoline stuff for cars any longer....

Hmmm... something's still not fully thought through here... ;-)

/sascha

Wonder if they were testing it over at Greenwich this morning... about 5 am?

Sure, water is a perfectly convenient form in which to store hydrogen if one can separate out the hydrogen as quickly as it is needed for the application.

The problem with storing hydrogen this way for automobiles is that water ice is not a particularly convenient way to store hydrogen for quick use. In much of North America, Europe, and Asia, water is mostly frozen almost a quarter of the year. If you're going to drive a water-powered car in New York, Chicago, Kansas City, Oslo, Moscow, Vladivostok... in January, either some toxic antifreeze needs to be added to your chemical slosh of water and aluminum/gallium alloy or you need to use some energy to heat your tank before and while driving.

I love the idea of just hopping over to my neighborhood nuclear plant to recycle my aluminum waste.

I'm sure the ecologists are going to love it too.

So when I go to my refueling station (formerly known as petrol station), instead of a pipe delivering a few gallons of liquid into a tank I'm going to have to first remove and then replace 350 lbs of solid pellets and some undisclosed quantity of water? Errr... are these places still going to be self-service or will they start employing teams of weightlifting staff? Where is Geoff Capes (http://en.wikipedia.org/wiki/Geoff_Capes) when we need him?

This looks very promising.

I hope someone takes this idea to the new auto "X Prize"

http://www.cbc.ca/technology/story/2007/04/02/auto-prize.html?ref=FreeEnergyNews.com

A few thoughts:

-Can you envision terrorists with water guns trying to sabotage our fuel supply?

-As someone else commented, the recycling plan for the alumina needs work.

-Will salt water work as well as fresh water?

"It's a simple matter to convert ordinary internal combustion engines to run on hydrogen," he says. "All you have to do is replace the gasoline fuel injector with a hydrogen injector."

Bollox. Gasoline (petrol to us Brits) is a liquid. Hydrogen is a gas. A cars fuel delivery system is designed to pump a liquid from tank to injector. The fuel lines and system are designed to be liquid leak-proof only.

Note to self - stay well away from this guys car when he starts it.

re: "Note to self - stay well away from this guys car when he starts it."

Given that he apparently managed to cause a violent explosion while doing the washing up, I'd say he was best avoided at all times...

Sascha seems to be forgetting that a power station is many times more efficient at producing energy than a car because it's done on such a large scale. By the time the technology has evolved and a distribution network established it will inevitably be cheaper. The delivery vehicles can also return the waste alumina instead of driving back empty - another win.

If you factor in oil exploration, drilling, transporting, refining and everything else oil is nowhere near an efficient alternative. It is just what we are familiar with. To allow existing cars a relatively simple route to indirectly running on electricity saves us all from replacing our cars and a huge leap forward environmentally.

Cost? We're paying significantly over £4 a gallon ($8) in the UK because gasoline is taxed so heavily: the $3 cost quoted doesn't scare me as much as the politician's take on it...

"Bollox. Gasoline (petrol to us Brits) is a liquid. Hydrogen is a gas. A cars fuel delivery system is designed to pump a liquid from tank to injector. The fuel lines and system are designed to be liquid leak-proof only."

Tell that to these guys then...

http://www.switch2hydrogen.com/h2.htm

As in the article, all they are doing on the injection side is replacing the gasoline/petrol injector with a hydrogen injector (whilst keeping the original injector intact to allow running on petrol also). I would assume the hydrogen fuel lines are their own and bypass the petrol lines.

Average American = 350 lbs - Ha! Sadly close to truth. Of course, if all y'all Brits could learn to cook a decent steak, among other things, you'd weigh more too... ;-)

Aluminium oxide - the 'waste' that is generated by the process is heavier

than the aluminium fuel itself. About twice as heavy, if my mental arithmatic

is correct. As this is formed by mixing with water also stored in the car, you

need to add 300+ lb of water to the weight of the car. It starts to add up...

2Al + 3H2O -> Al2O3 + 3H2

Also, all of this weight stays in the car. It doesn't get lighter as the fuel is

burnt.

Okay, YES this is a better way to "store" hydrogen then trying to compress and cool it. But weight issues aside, do you REALLY your car or anything else around you powered by something that gives off hydrogen whenever it is exposed to moisture? We'd need to store and transport the fuel in vacuum, and gods forbid there's ever an accident involving a car rupturing a tank and hitting a river or lake, instant boom.

<offtopic>

"Of course, if all y'all Brits could learn to cook a decent steak, among other things, you'd weigh more too... ;-)"

Sadly not very true this, it's mostly sugar and fat that explodes people to indecent sizes --- check the contents, a good steak is lean and not that high in calories.

Or you mean steak "drowned in some nasty gravy and obliterated by a mayonnaise-based glop"? Starting from an extra-fat miscut piece of meat? Then I misunderstood the word "decent"; tsk english/american false friends.

</offtopic>

"gods forbid there's ever an accident involving a car rupturing a tank and hitting a river or lake, instant boom."

No, hydrogen doesn't spontaneously combust in air. Liquid hyrgoen and liquid oxygen, yes - but even Michigan in February doesn't get cold enough to liquify oxygen.

Also, regarding the need to heat the fuel tank in colder climes - most cars sold in those areas already have cooling system heaters so the radiator won't freeze and burst. The tank doesn't have to be *really* warm, it just needs to stay above freezing. That's easy.

I think you need to re-read your texts.

It's not possible for the contents of the fuel tank to be more massive than at the beginning of the journey. Some of the mass in the tank has been converted to hydrogen that has been burned in the engine as fuel in order to move the car.

It would violate the known laws of physics for the content of the tank to be greater in mass after the 'fuel' has been used.

...as opposed to the wonderful safety inherent in volatile hydrocarbons...

The numbers are not there for hydrogen-fueled vehicles to achieve the efficiency of gasoline-powered vehicles, and this technology does not change that equation. Oil is expensive to extract from the ground, but it's really cheap to burn it. Water is relatively plentiful, but it's much more expensive to convert it to hydrogen, and that requires some energy source other than hydrogen. If you think renewable electricity sources are anywhere near ready to provide enough electrical power to drive this process for automobile fuel at anywhere near the cost of gasoline even with oil at say $100/barrel, you're not doing the math. The infrastructure we already have (and therefore don't need to pay to build) for generating electricity is largely based on fossil fuels, so if you suddenly put millions of electrolysis-powered cars on the road today we'd have no choice but to burn more coal and natural gas, which is a great way to accelerate global warming but will do nothing to get us to a sustainable transportation infrastructure. If we want to use electricity as a fuel for personal vehicles, we either have to 1) wait until the renewable (or nuclear - yikes!) capacity is there, 2) pray for some miraculous new technology and hope it won't take too long to deploy, or 3) burn more fossil fuels.

At some point people need to begin to accept the fact that after peak oil (and peak coal, and peak natural gas) we'll have no choice but to reduce our total energy consumption, and that means less driving and more public transportation. Either that or they need to produce the numbers showing we can power a personal-vehicle-heavy transportation infrastructure with renewable energy only. I've seen no such numbers from any analyst or researcher.

It takes 12.8 kW·h/kg (about 6 kW·h/lbs) to produce a pound of aluminum from aluminum oxide, the "exhaust" of the process these guys are promoting. yet they only get back 2 kW·h/lbs by "burning" it with water. So it only makes the energy crisis worse, not better.

How about Sodium??

Just refine some from a bit of salt (plenty in ocean!) and add some water. Lighter than Aluminum and probably easier to get. The waste product (NaOH) can clean your drains (or make soap) if needed. The reaction is well known see (http://www.theodoregray.com/PeriodicTable/Stories/011.2/) for more information.

Perpetual Cyclist needs to re-read the texts ...

During the journey, water is added to the 'fuel' tank though admittedly this is also carried onboard. In practice, the hydrogen is the lightest by far off all the elements involved so even including the water, the weight loss will be minimal.

Of course, there is another minor detail ...

... the hydrogen is reacted with oxygen to make water, and the oxygen is ingested from the atmosphere, and the water could be recovered for use in the aluminium reactor. Hence it actually would be possible for the overall weight to go up during a journey by the amount of oxygen consumed from the atmosphere.

How do they get the stuff to react fast enough to produce the neccessary volume of hydrogen? Even dropping bits of aluminium in acid doesnt do much more than bubble gently, due to aluminium's inconvenient oxide layer. You would need to optimise the surface area of the fuel pellets, which would mean costly machining of a resource which is going to be ultimately destroyed. You then need to manage the rate to ensure it continues to produce hydrogen fast enough to supply demand, but not so fast that an excess builds up, while having potential to supply more when needed, like when you need to climb a hill. Thats going to be some fiddly mechanics quite seperate from the engine, adding to the mass of the car.

Its also not that efficient in terms of mass of fuel to energy output. You carry 175kg of Al, and 350Kg of water to react with it, and produce 505.6Kg of Al(OH)3 and a measly 19.4kg of Hydrogen...

Whereas every gram of petrol carried is burned, at once releasing its stored energy, and reducing the mass of the car as the waste products are both gases, the hydrogen generator only release energy from the hydrogen produced. The Aluminium hydroxide is just additional, useless mass.

19.4kg of Hydrogen will release 7.8 x10^6 kJ of energy.

525Kg of Octane will release 2.5 x 10^7 kJ of energy, and you wont be carting round 95% of the original mass... Not really a contest.

Dont get me wrong, I think the idea has potential, but a portable hydrogen generator is not as convenient as it sounds, otherwise it would have already been adopted. The chemistry behind it is nothing new.

Tom:

Sodium cannot easily be extracted from sea water. Simply evapourating the water will leave you with salt, which is useless for this purpose.

Electrolysis of seawater will produce sodium hydroxide, which is not much use here either.

Electrolysis of molten NaCl (salt) will produce pure sodium, but what then. The stuff is very reactive, oxidises damn fast, and is generally pretty unpleasant.

Anyway, as electrolysis is the same process that is used to extract aluminium, it makes no sense to use sodium, as the cost increases to get the more reactive metal.

Aluminium is a good choice, as it is so abundant. It is unfortunately very hard to extract, hence why it isnt cheap.

Zinc is anther option, and is certainly easier to get, as you can extract it by a displacement reaction using carbon. This eliminates the need for costly electricity, but does mean you start producing nasty polluting gases like carbon dioxide and sulphur dioxide (depending on the zinc ore used). The resulting slower rate of reaction in the hydrogen generator could be improved by acidifying the water (perhaps using the nasty polluting gases. Now theres an idea...)

Of course the most effective solution would be to use your 525Kg to install a decent hydrogen containment system. You would be able to carry much more than 19.4Kg, thus giving you a far greater fuel economy than petrol. Perhaps the hydrogen generator could be a stationary device, in refueling (petrol) stations, or even in people's garages. You just use them to fill up your car at home, and the mass of other products doesnt matter much, as you simply get it collected and recycled when you get another load of fuel pellets delivered...

PS: Why isnt there an edit function on these boards?

Using aluminum as fuel is ridiculous (and the professor compares himself to Einstein, like a typical nacissistic crank).

At $1 per pound, how does he get that 350 pounds of fuel would cost $60? Aluminum is expensive because converting alumina to metal requires very high temperatures and a lot of electricity. Recycling your pellets would not cost less than the industrial cost of making aluminum. If anything it would cost more to deal with the gallium and form pellets, etc.

This proposal makes no sense in terms of energy or interms of the cost and logistics of dealing with heavy solid fuel pellets.

Only been suggested for iron filings so far injected into a normal internal combustion engine and sparked as usual - but I'm sure aluminium would give you that extra boost :-)

http://www.mng.org.uk/green_house/renewable_energy/metal_NS_article.htm

This came up (of all places) on a musicians message page. I've been trying to make sense of the numbers, but my calculations using basic chemistry was that to equal the energy in one US gallon (it was an American site) that you need about 8.1Kg of aluminium. Or for a 20 US gallon tank you need 162Kg of aluminium - pretty close to the 350 pound the article talks about (and a similar amount of water is required). At $1 per pound, that's $350 for a "tankful". That's against $60 for a 20 US gallon tankful of gasoline. That means that using the hydrogen would have to be almost 6 times more efficient to be competitive (and even fuel cells aren't 6 times more efficient than and IC engine - maybe three times).

As has been pointed out by somebody the vast majority of this has to be lugged around all the time as only the hydrogen gets lost (less than 5% of the mass). And for the guy who says it is win-win as you have an empty truck going back, I think I ought to point out that the mass moved to the refueling stop (excluding the water) is close to three times higher than for the petrol/gasoline and that driving back a full truck of the waste is a lot less fuel efficient than and empty truck or tanker.

In fact this isn't a completely new idea; there was something similar mooted in the 1980s, but with completely

http://www.keelynet.com/energy/cornish.htm

From what I can see the cycle to hydrogen loses something like 70% of the electricity required to make the aluminium, and that's excluding the distribution costs of all that extra mass to move around. That's not to mention all the extra fuel and waste mass in the vehicle which is bound to damage mileage (plus the weight of the in-car converter).

In short, I don't see how this will be cost effective.

"'gods forbid there's ever an accident involving a car rupturing a tank and hitting a river or lake, instant boom.'

No, hydrogen doesn't spontaneously combust in air. Liquid hyrgoen and liquid oxygen, yes - but even Michigan in February doesn't get cold enough to liquify oxygen."

You actually need an input of energy to get the H2 and the O2 to react. Usually an electrical discharge or a small flame, hence the danger of storing your compressed hydrogen fueled vehicle in your garage where the furnace and/or water heater usually live.

A patent for this was granted in 1982 and was assigned to IBM:

http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PALL&p=1&u=%2Fnetahtml%2FPTO%2Fsrchnum.htm&r=1&f=G&l=50&s1=4,358,291.PN.&OS=PN/4,358,291&RS=PN/4,358,291

A further patent was granted to IBM in 1988 for the production of the alloy:

http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&p=1&u=%2Fnetahtml%2Fsearch-adv.htm&r=2&f=G&l=50&d=PALL&S1=4358291.UREF.&OS=ref/4358291&RS=REF/4358291

Professor Woodall participated in both patents but not as first listed name.

It is presumably the latter patent's imminent expiry (they last 20 years don't they?) that has resulted in this old idea being pulled from its rocking chair and being paraded publicly in a short skirt and gaudy makeup. (Ugh!) Still doesn't make it a sensible idea for a car though, as Phil and others have correctly shown above...

The process of converting a Petrol engime to Hydrogen is a simple one as both are injected into the engine as a vapour (ie. a gas). In some cases, the pitson heads and the gasket may need to be reinforced as Hydrogen explodes slightly more violently than Petrol. This is exactly the same logic and reasoning behind why you should never put Petrol in a Diesel engine.

Also it's very easy to create a 'violent poof' with Aluminium: Simply wipe a sheet of Aluminium with a piece of cotton wool soaked in Silver Nitrate. Then submerse in water to create a 'violent poof' of steam. Add a catalyst like Gallium to the mix and it will just be more violent (as the Aluminium will heat up much faster). Yay for KS3 Chemistry!

Also the beauty of alloys over compounds is that another metal can (in most cases) be added to the mix without affecting the desired result. Thus trace amounts of Iron could be added, which would allow the pellets to be moved via magnetics (allowing easy removal of spent pellets).

It has also been mentioned that making pellets and adding Gallium would push up the price of recycling the Alumina (or Alum). Since the Gallium is untouched by the entire process, there would almost always be the required amount available, and replacing any that has somehow been 'spoilt' would be a relatively cheap affair. As for making pellets, Aluminium and Gallium are both fairly easy to shape and press, so pellets are not a problem. Of course the pellets would need to be vacuum sealed before being used to create the hydrogen, otherwise they would react with the air and become effectively useless. Although perhaps a Silver Nitrate -based lubricant could be developed for the delivery mechanism.

As to the weight problem I see two possible solutions. a) Phil's idea of a static Hydrogen generator producing fuel at home, or b) just drop the pellets by the side of the road, once they're spent and a road-sweeper could collect them at regular intervals, This would be pretty simple with added iron :P

Hmm, this reminds me of the farce of the the very poor missunderstood physics expoused in the mainstream public media of the so called liquid explosive , from the official claptrap rubbish which enable the British Police to terrorise all the innocent aviation travellers in the UK heading both east and west(very convenient one could for the armed police to renew their sagging and somewhat adverse public image of shooting many down innocent victims in public places which they appear to do on a somewhat regular basis and get away with literal murder without any punishment of the offending police officers other than a notional ha ha reprimand "Nelson style"(you or I doing the same identical crime would be looking at a very long term prison sentence , nice to have an official police badge and you can get away with murder))

One question , who failed to comprehend basic laws of physics, thermodynamics and the laws of diminishing returns for this crap idea ?

Or how soon we forget the enviromentally friendly Stanley or White Steamer of the pre WW2 era?(could be fueled by non polluting liquid vegetable oil(and even pre heat it as well) and the lower external combustion temperature produce minimal enviromental pollution in comparison to todays so called clean lean burning fossil fueled motor vehicles we drive today!

Oh well , back to putting one's head in the sand like a good ostrich , let the blind lead us to oblivion Tony Blair style!

"Bollox. Gasoline (petrol to us Brits) is a liquid. Hydrogen is a gas. A cars fuel delivery system is designed to pump a liquid from tank to injector. The fuel lines and system are designed to be liquid leak-proof only."

Tell that to these guys then...

http://www.switch2hydrogen.com/h2.htm

As in the article, all they are doing on the injection side is replacing the gasoline/petrol injector with a hydrogen injector (whilst keeping the original injector intact to allow running on petrol also). I would assume the hydrogen fuel lines are their own and bypass the petrol lines.

In which case you're not "just replacing the injectors". Don't get me wrong, I don't really know about the underlying technology. But I do know that trying to sell this system as something you can magically retrofit to *exisiting* automotive technology is a little disingenous.

It's the same as automotive LPG. Great idea. But the conversion costs prohibit a large scale-uptake. Any "green intiative" for cars will have to understand the fact that unless *exisiting* cars can use it, it's a dead duck. I'd love to be green, I really would. If I can do it byt putting a different fuel in my car - brilliant. I'd even accept a slight performance hit. But I'm not going to buy a new car.

Everyone keeps going on about the weight of water required. Why?!?!? Don't you realise that water is the sole product of burning hydrogen? Recycle it, dagnabbit!

Since you are splitting water into oxygen and hydrogen with this process, and exhausting the hydrogen via the engine, the overall weight of the car must go up, as hydrogen has a negative weight, as proven by the fact that hydrogen balloons go up.

But for the conspiracy part ...

IBM have apparently held the patent on this process for years, and as we all know, water-powered cars have only been stopped by the oil companies and big business allies to protect the oil profits! So now that oil is getting close to running out, they are finally releasing details of the process that has been secretly hidden in the US patent office for years!

I own a diesel car which is much more ecomonical than a petrol 1. Most Diesel cars can run on vegetable oil without any modification or with very little modification. Why aren't we allowed to run our cars on this renewable fuel as a stopgap measure?

But to collect the "waste" water will require many more bits which will add to the overall weight to be lugged around.....

And please tell me that Phil above does NOT live in London. Hydrogen generator in the garage (presumable with storage). Sheesh !! One good ciggie and, literally, BANG goes the neighbourhood !!

because most folk are afraid to go any where near the idea of a self perpetuating machine.

This isn't a perpetual machine but the very thought of self refuelling, IE in simplistic terms water-hydrogen-water, seems to scare folks.

Probably because there's no money to be made.

You cant just recycle the water. Yes, you could use the water you make to supplement the on-board store, but since out of every water molecule, only one hydrogen atom goes to make hydrogen gas. The oxygen and the other hydrogen react with the aluminium to make aluminium hydroxide. Hence you would need at least half the total mass of water you need. I effectively ignored the combustion side of the reaction in my earlier post, as the oxygen needed and the waste water produced both dont affect the mass of the car (coming from and returning to the atmosphere).

Ishkandar: I dont live in London, but this neednt be a problem. At least no more of a problem than storing hydrogen in the car. The whole point of the hydrogen generator is that you generate hydrogen when you need it, otherwise you might as well use the electricity to electrolyse water in the first place, and not the aluminium ore, and deliver hydrogen gas direct to the customer. Its far more efficient that way.

In my experience hydrogen liberation by adding metal to acid (or water, if the metal is reactive enough) is a pretty quick affair. You wouldnt need to limit the rate, so could just chuck a load of pellets into the generator shortly before you go, and bingo, you're tank will soon be full... (the reaction is very fast at the start as there are lots of unreacted particles of metal and water mixed together, so the chance of them meeting and reacting is initially very high).

To the diesel car owner: The reason you cant run your car on diesel is because you would not be paying fuel tax, and our dear (soon to be leader) chancellor takes a dim view of people sidestepping his money harvesting schemes... Theres no reason why you cant do it on the sly, except that your car smells like a chippy when you are driving along...

also recycling the water would actually increase the mass of the car as you drive. You would be adding the mass of the oxygen from the air used to burn the hydrogen in the engine.

This lack of edit function is really annoying...

basic physics, Mr Rice. A vapour is a liquid suspended in a gas.

A gas is a gas.

...

Although vapour and gas are frequently (incorrectly) used interchangeably, vapor refers to a gas phase in a state of equilibrium with identical matter in a liquid or solid state below its boiling point, or at least capable of forming solid or liquid at the temperature of the vapor.

Erm... No big tanks of water required...

new pellets + water = Hydrogen and old pellets

Hydrogen + oxygen = water and vroom.

take the exhaust water and add it back to start again.

That's assuming that this whole idea isn't just a poof of hot air from another research looking for a funding cheque on the next bandwagon.

As to diesel engines using veggie oil - They were designed by Mr Diesel to run on peanut oil. Veggie oil from the supermarket / wholesaler / fast food purveyor still needs to have "fuel" tax paid on it before you burn it making it only marginally less expensive than the pump. Also, more particulate emissions and the constant hanger for fried food after catching a whiff of the exhaust gases.

Vegetable oil may become an important transitional fuel because it's *relatively* carbon-neutral. But there is *no way* we can produce enough vegetable oil to fuel vehicles at the current rate of vehicle use without *dramatically* increasing the cost of food (the land we need for growing that large a crop of fuel-producing vegetables is currently being used for food), which will greatly reduce food availability in the developing world, leading to starvation. It's already happening here in North America as corn prices, driven up by ramped-up ethanol production, are making corn products very difficult to afford for poorer Mexicans. I for one am not willing to starve people so I can drive a personal vehicle, but I can't speak for the masses of suburban commuters.

In this time of anxiety lots of cranks are going to offer us "new" technologies that give us energy for nothing. It's a dangerous con game, and physics will not oblige.

Quote: "Erm... No big tanks of water required...

new pellets + water = Hydrogen and old pellets

Hydrogen + oxygen = water and vroom.

take the exhaust water and add it back to start again."

Its not quite that simple. There are two separate reactions happening. First you are liberating hydrogen from the water using aluminium.

2Al + 6H2O --> 2Al(OH)3 + 3H2 (subscript is also missing from these boards...)

Notice 6 water molecules give only 3 hydrogen molecules (6 hydrogen atoms). All 6 oxygens and the other 6 hydrogens are locked up in the aluminium hydroxide (and thus are not available for 'recycling').

In the engine an entirely seperate reaction takes place.

3H2 + 1.5O2 --> 3H2O

The oxygen in this reaction comes from the air, and is totally separate from the oxygen that the hydrogen was originally bonded to. That oxygen is now part of the aluminium hydroxide.

We could collect our new water and use that to react with more aluminium pellets. In fact we would produce 175Kg of water this way (19.4kg of hydrogen burns to make 175Kg of water - the extra 155.6Kg coming from the mass of oxygen taken from the air), so we would only need to start out with 175Kg. We would still end up with 525Kg of aluminium hydroxide, so our car would end up carrying more mass than when it started.