This is the second part of my posts about my the undergraduate physics research I participated in this summer. Check out part 1 here. When we left off I had received a space grant (GRANTS… IN… SPAAAAACE) that would pay me for an internship position at my school. I’m not going to lay out my two months of research as a chronological story because that seems rather boring to me. Instead, I’m going to give some background on the subject of my research, some of the cooler stuff I got to do, and what we found out. So, as I said in part 1, the scientists I was working with specialized in materials science, specifically hydrogen storage.
So, why do we care about storing hydrogen? Well, hydrogen can be used as a fuel for vehicles or for generating power. Basically we can store energy in the form of hydrogen and then burn it without releasing any greenhouse gases, just water. Greenhouse gases are bad. Actually, water vapor is technically a greenhouse gas but not one we’re as worried about as carbon dioxide. Also, we don’t really burn it, we use a hydrogen fuel cell to turn the hydrogen directly into electrical energy.
There’s a problem though. While hydrogen is a great energy carrier with a greater efficiency than gasoline, it has to be stored either under high pressure (5,000 – 10,000 psi) or at extremely low temperatures (lower than -252.8ºC). That really makes it not ideal as a fuel, especially in vehicle applications. High pressure tanks are heavy and that added weight reduces the overall efficiency of a hydrogen fuel cell vehicle. Plus, people get a little worried about having an explosive gas under high pressure in their car (even though gasoline is probably actually way more dangerous than hydrogen in that respect). Everyone starts thinking about the Hindenburg. Also, a pressurized tank has to be a certain shape, and you can’t just change that shape to fit it in nicely to the car like you can a gasoline tank. Cooling hydrogen down takes a lot of extra cooling equipment and insulation that also isn’t ideal.
So, storing hydrogen by itself has a lot of drawbacks. But, fear not, there are other ways we can store hydrogen. Materials hydrogen storage is a way to take those hydrogen atoms which want to shoot off into the atmosphere or turn into fireballs and bind them to some material in a solid state. Then, you just reverse that process to get the hydrogen back out. I say “just”, but it’s rarely this simple and sometimes the process is not reversible in any practical way.
One type of materials storage for hydrogen is complex chemical storage. This is where hydrogen is bound up in a molecule. This has some advantages in that you have a higher weight percentage of hydrogen in these molecules and many of them are stable at moderate pressures and temperatures (so no crazy heavy pressure or cryo tanks). This storage method has some drawbacks too. One is that because hydrogen is so tightly bound that it might require crazy high temperatures to get the hydrogen out (400ºC or higher) and some of these molecules are highly reactive with things like air and water.
To get a bit more specific, the material I was working with is called sodium alanate or sodium aluminum hydride (NaAlH4). A bunch of research has been done on this stuff since the early 2000’s because it seems promising as a way to store hydrogen and later release it. Now, if you watched the video I linked to in the previous paragraph, you may be asking, “Chris, why don’t you just mix the stuff with water and use that reaction to power your vehicle or power plant?” Good question reader. You actually could do that, and you would generate heat which you could use to boil water and turn a turbine of some kind. Mixing sodium aluminum hydride with water is a very violent reaction that’s extremely hard to control (but then again so is gasoline exploding). Believe me, I cleaned out a mixing vessel that had a light coating of NaAlH4 in it with water and a foot high jet of flame shot out of it. But, remember, we’re talking about hydrogen storage. We want this material to store hydrogen, release it, and then absorb it again over and over. The water and sodium alanate reaction is non-reversible.
The release of hydrogen, called dehydrogenation or desorption begins with the following two reactions:
(There’s a third reaction where sodium lets go of hydrogen, but it occurs at temperatures that are too high for us to care about).
So, we want to heat up this material (usually a powder) so that it lets go of its hydrogen atoms and then use those atoms to power a fuel cell. We have two problems here. First, the temperature we have to heat it to is too high. Sodium Aluminum Hydride at atmospheric pressure initially begins to let go of its hydrogen atom at around 180° C. Producing the heat to get this reaction started takes energy, and using this energy reduces the overall efficiency of the system. Second, this hydride releases the hydrogen pretty slowly. At the minimum temperatures to begin releasing hydrogen it can take hours for all of the hydrogen to be released.
Luckily, many scientists have already figured out a way to reduce both the temperature and the time it takes for these dehydrogenation reactions to occur. First, they’ve found out that they can reduce the desorption/dehydrogenation temperature of this (and other) hydride(s) by doping them with catalysts. What the hell does that mean? Well, basically you take another material, like titanium trichloride (TiCl3) and mix a small amount of it (less than 10% by weight). This reduces the temperature at which hydrogen is released. How does it do this? That’s actually an unanswered question. There have been studies that have come to opposite conclusions about how dopants affect the desorption kinetics (how fast and at what temperature the release hydrogen) of hydrides. Science doesn’t always require that you start with a theory and then test it. Sometimes you can just do something and see what happens. In this case the why may be secondary to the what. What type and amount of dopant gives the best outcome (lowest desorption temperature) for sodium aluminum hydride?
What about the second problem of the reaction taking too long? Again, previous scientific studies have shown that if you mill the dry powder form of sodium aluminum hydride you get faster dehydrogenation reactions. Milling is basically taking the dry powder, sealing it in a jar with some ball bearings and then putting it into what is essentially a really high power paint shaker. A science shaker.
By the way, because sodium aluminum hydride reacts with oxygen and the moisture in the atmosphere it has to be kept in a dry nitrogen atmosphere. In my case it was kept in a glove box, which looks cool, but is a major pain to work with. Here’s a glovebox at the Advanced Photon Source at Argonne National Lab.
The glovebox at Rowan was mostly metal, which was better for long term storage, but was much harder to work with. Who knew science could be so hard on your back?
So, all of that preceding was background on the subject. My actual research built on the work of those previous studies I mentioned and on the work of my advisor and her previous research students. What we were trying to figure out was; what happens to the structure of sodium aluminum hydride when we add these dopants and/or mill it, and how does this relate to how fast the hydride releases hydrogen.
There’s a few ways we can study the small-scale structure of materials. First we can use scanning electron microscopy (SEM). SEM is a lot like regular microscopy except that we use electrons instead of photons to probe smaller structures of materials. This works because electrons have a smaller wavelength than visible light. Yes, electrons have wavelengths. Quantum physics y’all.
Here’s a gif of a series of SEM images of sodium aluminum hydride.
At the highest magnification, we’re looking at features on the order of nanometers. While this seem really small, it actually only probes the gross structure of this hydride. What did the SEM tell me? Well, the ball milled NaAlH4 is clumpier, an assuredly unscientific word, but if I show you a picture of the milled and un-milled NaAlH4 side-by-side, you can see what I mean.
As far as we know, these are the same materials, at the exact same magnification. The difference is pretty striking, huh? Why would this difference cause the hydrogen to desorb faster from the second powder? Well, that’s an open question, and one that we weren’t able to answer over the summer.
Now, you’ll remember that I said that SEM lets us look at the gross or large-scale structure of materials. Then how do we probe smaller structures? That’s where X-Rays come in. At Rowan we have a laboratory scale X-ray Diffraction machine. I actually used this device before I did any SEM scans. It’s a really neat machine.
I’m not going to go deep into the physics of it, but X-rays allow us to probe the atomic crystal lattice structure of materials.
Basically if there’s a regularly ordered crystal structure in your material then at certain angles the X-rays will constructively interfere with each other and produce peaks of high intensity in the detector, like so:
These graphs actually come from data that we gathered at Argonne National Lab at this place,
Using this detector,
The major difference between our lab scale XRD and the one at APS is that the APS is way, way brighter, meaning it has way higher energies. It also has four different detectors that can measure a wider range of angles of scattering from the beam. This can give us different kinds of information about the structure of the materials. More on that later.
So at Argonne we took our powder and heated it up to 135° C and held it at that temperature for an hour, while scanning it with the X-ray detector periodically. Why that temperature? Well, for one thing it’s below the normal desorption temperature for NaAlH4, and it is also a temperature at which previous studies were conducted, so we could compare the new data directly to old data.
This is the “heating stage” that we put the sample powders in:
So check this out, we heated samples doped with different catalysts and measured the crystal structure of each. Here’s an XRD graph of a doped sample:
So, it looks pretty similar to the graph I post further up. The different colored lines represent the intensity of scattered electrons at various “d-spacings”. Basically, if there’s a peak, then that means there’s a repeating structure of atoms that is separated by that much distance in angstroms (0.0000000001 meters). The blue line is the scan taken at room temperature. Then going from the green to the orange, and then to the red line are measurement taken at 135º C over time. Okay, so look at the peaks I’ve marked with a dashed vertical line. See them? Those are known peaks for sodium aluminum hydride. We expect to see them because, as far as we know, our sample is crystalline NaAlH4. We’re interested in how these peaks change as the sample is continually heated. So far, so good. So check out those solid black vertical lines. Those lines indicate the presence of aluminum. That indicates that hydrogen desorption has begun. Now, go back up and look at the first XRD graph above. Do you see it? No aluminum lines! Oh dang. That means that we really have lowered the desorption temperature of our hydride. These lines are only present in the doped samples (there are more graphs of the other doped samples, but they’re very similar to the one above).
The third type of scan we did is called ultra-small angle X-ray scattering (USAXS). It uses the same beam as the XRD above, but with a different detector. This scan provides some more information about the non-crystal structure of the material. It’s based on a pretty complicated mathematical model that contains in it information about the size, shape, and fractality (if that’s a word) of populations of particles in the materials. A graph of USAXS data looks like this:
Okay, so that line that goes from upper left to to lower right, with the “hump” or “shoulder” near the lower right is the data obtained from the USAXS scan of our hydride. By applying the mathematical model known as “unified fit”, represented here by the magenta and green lines, and making several assumptions about the shape of the particles we can get an idea about the radius, and porosity of the particles in the material. Why is this important? Like I said before, we think that doping and milling cause structural changes to sodium aluminum hydride that lead to the effects on dehydrogenation that we see. USAXS is another way to look at structure. What did we learn from this data? Honestly, I’m not entirely sure. This model is way more complicated (to me) than XRD or SEM, so I have a bit of trouble wrapping my head around it. All I can say is that we need to do more research.
That’s pretty much it. This may be the longest thing I’ve written in a long time. I hope it held your attention. I even started to zone out writing it about half way through. I hope to put up more posts about my research when I start again in the fall, and about the sorts of things I’m learning in my classes in the fall.