Tag Archives: metal hydrides

Hydrogen transport in metallic membranes

Leave a reply

The main products of my company, REB Research, involve metallic membranes, often palladium-based, that provide 100% selective hydrogen filtering or long term hydrogen storage. One way to understand why these metallic membrane provide 100% selectivity has to do with the fact that metallic atoms are much bigger than hydrogen ions, with absolutely regular, small spaces between them that fit hydrogen and nothing else.

Palladium atoms are essentially spheres. In the metallic form, the atoms pack in an FCC structure (face-centered cubic) with a radius of, 1.375 Å. There is a cloud of free electrons that provide conductivity and heat transfer, but as far as the structure of the metal, there is only a tiny space of 0.426 Å between the atoms, see below. This hole is too small of any molecule, or any inert gas. In the gas phase hydrogen molecules are about 1.06 Å in diameter, and other molecules are bigger. Hydrogen atoms shrink when inside a metal, though, to 0.3 to 0.4 Å, just small enough to fit through the holes.

The reason that hydrogen shrinks has to do with its electron leaving to join palladium’s condition cloud. Hydrogen is usually put on the upper left of the periodic table because, in most cases, it behaves as a metal. Like a metal, it reacts with oxygen, and chlorine, forming stoichiometric compounds like H2O and HCl. It also behaves like a metal in that it alloys, non-stoichiometrically, with other metals. Not with all metals, but with many, Pd and the transition metals in particular. Metal atoms are a lot bigger than hydrogen so there is little metallic expansion on alloying. The hydrogen fits in the tiny spaces between atoms. I’ve previously written about hydrogen transport through transition metals (we provide membranes for this too).

No other atom or molecule fits in the tiny space between palladium atoms. Other atoms and molecules are bigger, 1.5Å or more in size. This is far too big to fit in a hole 0.426Å in diameter. The result is that palladium is basically 100% selective to hydrogen. Other metals are too, but palladium is particularly good in that it does not readily oxidize. We sometime sell transition metal membranes and sorbers, but typically coat the underlying metal with palladium.

We don’t typically sell products of pure palladium, by the way. Instead most of our products use, Pd-25%Ag or Pd-Cu. These alloys are slightly cheaper than pure Pd and more stable. Pd-25% silver is also slightly more permeable to hydrogen than pure Pd is — a win-win-win for the alloy.

Robert Buxbaum, January 22, 2023

A clever, sorption-based, hydrogen compressor

Leave a reply

Hydrogen-powered fuel cells provide weight and cost advantages over batteries, important e.g. for drones and extended range vehicles, but they require highly compressed hydrogen and it’s often a challenge compressing the hydrogen. A large-scale solution I like is pneumatic compression, e.g. this compressor. One would combine it with a membrane reactor hydrogen generator, to fill tanks for fuel cells. The problem is that this pump is somewhat complex, and would likely add air impurities to the hydrogen. I’d now like to describe a different, very clever hydrogen pump, one that suited to smaller outputs, but adds no impurities and and provides very high pressure. It operates by metallic hydride sorption at low temperature, followed by desorption at high temperature.

Hydride sorption -desorption pressures vs temperature.

Hydride sorption -desorption pressures vs temperature, from Dhinesh et al.

The metal hydriding reaction is M + nH2 <–> MH2n. Where M is a metal or metallic alloy and MH2n is the hydride. While most metals will undergo this reaction at some appropriate temperature and pressure, the materials of practical interest are exothermic hydrides that is hydrides that give off heat on hydriding. They also must undergo a nearly stoichiometric absorption or desorption reaction at reasonable temperatures and pressures. The plot at right presents the plateau pressure for hydrogen absorption/ desorption in several, exothermic metal hydrides. The most attractive of these are shown in the red box near the center. These sorb or desorb between 1 and 10 atmospheres and 25 and 100 °C.

In this plot, the slope of the sorption line is proportional to the heat of sorption. The most attractive materials for this pump are the ones in the box (or near) with a high slope to the line implying a high heat of sorption. A high heat of sorption means you can get very high compression without too much of a temperature swing.

To me, NaAlH4 appears to be the best of the materials. Though I have not built a pump yet with this material, I’d like to. It certainly serves as a good example for how the pump might work. The basic reaction is:

NaAl + 2H2 <–> NaAlH4

suggesting that each mol of NaAl material (50g) will absorb 2 mols of hydrogen (44.8 std liters). The sorption line for this reaction crosses the 1 atm horizontal line at about 30°C. This suggests that sorption will occur at 1 am and normal room temperature: 20-25°C. Assume the pump contains 100 g of NaAl (2.0 mols). Under ideal conditions, these 100g will 4 mols of hydrogen gas, about 90 liters. If this material in now heated to 226°C, it will desorb the hydrogen (more like 80%, 72 liters) at a pressure in excess of 100 atm, or 1500 psi. The pressure line extends beyond the graph, but the sense is that one could pressure in the neighborhood of 5000 psi or more: enough to use filling the high pressure tank of a hydrogen-based, fuel cell car.

The problem with this pump for larger volume H2 users is time. It will take 2-3 hours to cycle the sober, that is, to absorb hydrogen at low pressure, to heat the material to 226°C, to desorb the H2 and cycle back to low temperature. At a pump rate of 72 liters in 2-3 hours, this will not be an effective pump for a fuel-cell car. The output, 72 liters is only enough to generate 0.12kWh, perhaps enough for the tank of a fuel cell drone, or for augmenting the mpg of gasoline automobiles. If one is interested in these materials, my company, REB Research will be happy to manufacture some in research quantities (the prices blow are for materials cost, only I will charge significantly more for the manufactured product, and more yet if you want a heater/cooler system).

Properties of Metal Hydride materials; Dhanesh Chandra,* Wen-Ming Chien and Anjali Talekar, Material Matters, Volume 6 Article 2

Properties of Metal Hydride materials; Dhanesh Chandra,* Wen-Ming Chien and Anjali Talekar, Material Matters, Volume 6 Article 2

One could increase the output of a pump by using more sorbent, perhaps 10 kg distributed over 100 cells. With this much sorbent, you’ll pump 100 times faster, enough to take the output of a fairly large hydrogen generator, like this one from REB. I’m not sure you get economies of scale, though. With a mechanical pump, or the pneumatical pump,  you get an economy of scale: typically it costs 3 times as much for each 10 times increase in output. For the hydride pump, a ten times increase might cost 7-8 times as much. For this reason, the sorption pump lends itself to low volume applications. At high volume, you’re going to want a mechanical pump, perhaps with a getter to remove small amounts of air impurities.

Materials with sorption lines near the middle of the graph above are suited for long-term hydrogen storage. Uranium hydride is popular in the nuclear industry, though I have also provided Pd-coated niobium for this purpose. Materials whose graph appear at the far, lower left, titanium TiH2, can be used for permanent hydrogen removal (gettering). I have sold Pd-niobium screws for this application, and will be happy to provide other shapes and other materials, e.g. for reversible vacuum pumping from a fusion reactor.

Robert Buxbaum, May 26, 2017 (updated Apr. 4, 2022). 

What is the best hydrogen storage medium?

4 Replies

Answering best questions is always tricky since best depends on situation, but I’ll cover some hydrogen storage options here, and I’ll try to explain where our product options (cylinder gas purifiers and methanol-water reformers) fit in.

The most common laboratory option for hydrogen storage is inside a tank; typically this tank is made of steel, but it can be made of aluminum, fiberglass or carbon fiber. Tanks are the most convenient source for small volume users since they are instantly ready for delivery at any pressure up to the storage pressure; typically that’s 2000 psi (135 atm) though 10,000 (1350 atm) is available by special order. The maximum practical density for this storage is about 50 g/liter, but this density ignores the weight of the tank. The tank adds a factor of 20 or to the weight, making tanks a less-favored option for mobile users. Tanks also add significantly to the cost. They also tend to add impurities to the gas, and there’s a safety issue too: tanks sometimes fall over, and compressed gas can explode. For small-volume, non-mobile users, one can address safety by chaining up ones tank and adding a metal membrane hydrogen purifier; This is one of our main products.

Another approach is liquid hydrogen; The density of liquid hydrogen is higher than of gas, about 68 g/liter, and you don’t need as a tank that’s a big or heavy. One problem is that you have to keep the liquid quite cold, about 25 K. There are evaporative losses too, and if the vent should freeze shut you will get a massive explosion. This is the storage method preferred by large users, like NASA.

Moving on to metal hydrides. These are heavy and rather expensive but they are safer than the two previous options. To extract hydrogen from a metal hydride bed the entire hydride bed has to be heated, and this adds complexity. To refill the bed, it generally has to be cooled, and this too adds complexity. Generally, you need a source of moderately high pressure, clean, dry hydrogen to recharge a bed. You can get this from either an electrolysis generator, with a metal membrane hydrogen purifier, or by generating the hydrogen from methanol using one of our membrane reactor hydrogen generators.

Borohydrides are similar to metal hydrides, but they can flow. Sorry to say, they are more expensive than normal metal hydrides and they can not be regenerated.They are ideal for some military use

And now finally, chemical materials: water, methanol, and ammonia. Chemical compounds are a lot cheaper than metal hydrides or metal borohydrides, and tend to be far more readily available and transportable being much lighter in weight. Water and/or methanol contains 110 gm of H2/liter;  ammonia contains 120 gms/liter, and the tanks are far lighter and cheaper too. Polyethylene jugs weighing a few ounces suffices to transport gallon quantities of water or methanol and, while not quite as light, relatively cheap metallic containers suffice to hold and transport ammonia.

The optimum choice of chemical storage varies with application and customer need. Water is the safest option, but it can freeze in the cold, and it does not contain its own chemical energy. The energy to split the water has to come externally, typically from electricity via electrolysis. This makes water impractical for mobile applications. Also, the hydrogen generated from water electrolysis tends to be impure, a problem for hydrogen that is intended for storage or chemical manufacture. Still, there is a big advantage to forming hydrogen from something that is completely non-toxic, non-flammable, and readily available, and water definitely has a place among the production options.

Methanol contains its own chemical energy, so hydrogen can be generated by heating alone (with a catalyst), but it is toxic to drink and it is flammable. I’ve found a  my unique way of making hydrogen from methanol-water using  a membrane reactor. Go to my site for sales and other essays.

Finally, ammonia provides it’s own chemical energy like methanol, and is flammable, like methanol; we can convert it to hydrogen with our membrane reactors like we can methanol, but ammonia is far more toxic than methanol, possessing the power to kill with both its vapors and in liquid form. We’ve made ammonia reformers, but prefer methanol.