![[Go to Home page]](../images/smallseg.gif){kind=small} |
Australian Academy of Science Science at the Shine Dome Canberra, 3-5 May 2006 |
|
Full listing of papers
|
SCIENCE AT THE SHINE DOME ANNUAL SYMPOSIUM Science on the way to the hydrogen economy 5 May 2006
Hydrogen storage: Status and prospects
(Click on image for a larger version) Thank you, Mr Chairman, and thank you to the Academy for the opportunity to talk about hydrogen storage. You have heard that hydrogen storage is on the critical pathway to the Hydrogen Economy. Indeed it turns out to be a tough nut to crack. So I am going to talk to you about the likelihood that we might have the ‘right hammer’ to achieve this at some time soon. There are some powerful forces at work here – contests of all sorts are going on at the global scale. Shown here are some of the factors that are driving such research worldwide. These are, of course, largely political factors. Something that has been important to us is the announcement of the project to build the Freedom Car, which contains a large amount of money. This very effective incentive encourages research in this direction, as all researchers know. Some of the problems that occur along the way to developing such projects, firstly involve hydrogen storage, particularly for automobiles. There are, of course, other requirements for hydrogen fuel, but I am going to talk mostly about automobiles because that, in many senses, is the most difficult part of the problem. Some targets have been established and I am going to talk about these things in more detail, shortly. I will particularly cover how these targets are hard to meet. Even though, as you heard, these sound like easy targets compared to the performance of gasoline, the reality is that the world’s automotive industry has developed on the basis of the availability of petroleum products, and we now somehow must compete with those in technical and economic ways. Another reality is that, even though 30 or 35 years ago there was a lot of hope for metal hydrogen systems, which is where I came into this line of business some 20 years ago, and nickel metal hydrogen batteries emerged along the way (and they are very good), nevertheless, these developments haven’t produced the larger-scale, lighter-weight storage systems that we need for cars. The only technologies that presently serve the purpose may only be applied in interim ways, though they are presently available. The ones that you have heard about include cryogenic storage and pressurised gas tanks. I am now going to talk quite a bit about this example, because I want to try and convince you that what we really need is condensed matter systems. As the last dot point of this slide, note that some alternatives are presented. Previous speakers have given you some sort of an introduction to the concepts here, so I won’t repeat myself, on the well-known principle among university teachers that, if you tell a whole lot of first-years a whole lot of information in first year, they’ll remember it all when they come back in second year! I hope the students who are present take that on board when they go to university. So you have heard something about some of these. I won’t talk very much about ‘Chemical’, by this I mean - fuels which contain hydrogen but which do not directly supply hydrogen. This can also mean those chemical systems which can not be directly recharged with hydrogen on board the vehicle and so this must be done elsewhere. An example of this is gasoline. It contains hydrogen and you burn it, but you have to get the carbon dioxide back by some other means and take it round the cycle again, and that of course, is how we got to where we are: trying to find a way out of this loop. The real nature of the problem is shown in this table. The mass density of hydrogen energy per kilogram is very high – it is excellent, compared with that of gasoline. Unfortunately, the volume density is lousy. If you were to try and compete with gasoline, then you would end up needing about 35 wt% hydrogen (120 MJ/kg for hydrogen compared to 44 MJ/kg for petrol) at about four times liquid density (8 MJ/litre for hydrogen compared to 32 MJ/litre for petrol). So this is, in terms of a level playing field, the kind of competition between technologies that needs to go on. You have heard and seen lots of hydrogen-fueled vehicles in images. I am interested in the numbers of ‘70 MPa/700 bar/10,000 PSI’. I don’t know how many of you have ever played with pressurised hydrogen at 70 MPa/700 bar – roughly 700 atmospheres – or 10,000 PSI, in ‘old money’. Certainly in my lab we work with pressures higher than this, and I can assure you that we get rather respectful once we get to that kind of pressure. The concept of having a substantial fraction of a cubic metre of 700 bar hydrogen in the boot or trunk of my car isn’t really something that fills me with great joy, I have to say. However, this is the current reality. These are the most advanced pressurised gas storage systems that have so far been approved. You have already seen this image a couple of times today. This slide particularly shows what is behind the plume of water coming off: the hydrogen tanks in the roof of the bus, the fuel cell and so on. I am interested in the figures of ‘350 bar (5000 psi)’. You may notice that the volume of the fuel storage is rather high – that is the problem with pressurised gas. That turns out to be the real problem, if you like. How is performance measured? The US Department of Energy has established some goals. They have been revised a few times but this is, roughly speaking, the current set. They are hard goals to meet but they are goals, as you have heard, which have some relation to the performance of gasoline. It has also been explained that there is the need to rely on the potential extra efficiency of fuel cells relative to the roughly 20 per cent efficiency of an internal combustion engine, to make up the difference in energy storage density. Now, 2005 has gone by and I will tell you now that these goals really have not been met. 2010 is coming up, and so I am going to talk to you about the likelihood of meeting the 2010 and then the first of the 2015 goals. There are other goals too, but the focus is very much on energy density. It is, if you like, a little way down the track for determining how the other goals might be tackled and met. So here is a diagram, which I think originally is due to Louis Schlapbach in its concept and Andreas Züttel, and which allows a comparison of storage methods. The lines are drawn in blue for 2005, purple for 2010, and red for 2015 to show the DoE goals. So we are interested in moving, by 2015, somewhere up toward the box area at the top right. Liquid hydrogen is situated here in the diagram; note that it is not actually dense enough by volume to meet the 2015 target. Now, these goals are debatable and there are plenty of companies and perhaps even countries which do not accept these goals, Nevertheless, they represent some kind of reality against which to measure performance, and just how long they will serve as a benchmark is unknown. I draw your attention to this curve down here near the origin, and I need to explain briefly what this curve represents. In a minute I will also show you the model by which one arrives at such a curve, but this is the performance of my simulation of the current quantum technologies tank. This is based on the properties announced by the DoE in their annual review in 2005. This particular tank has just fairly recently been approved for 700 bar usage in Europe. This storage mechanism runs at 700 bar, and you can see its performance. In 2005 it was, as we say, within a “bull’s roar” of the 2005 target and it was just under 4 wt% and, I think, 24 kg per cubic metre. So somehow it is necessary to get from this area to the top right of the diagram. There you can see the materials that were of interest – indeed they are still of interest in a scientific context, and they are analogues for the metal hydrogen batteries that have been studied since the 1970s. Then situated out at the top far right, there are some materials that I will talk briefly about, towards the end of the presentation. Here are the liquid carbons that are readily available and that we are really competing with, in many senses. So, high-pressure gas storage? This is a very simple model and mechanical engineers will recognise these formulae as coming from AS1210, the Australian Standard for the design of pressure vessels. The system comprises a cylindrical container with spherical ends, and the strength is limited by the cylinder because cylinders are of lower symmetry than spheres and less strong. It is necessary to know the strength of the material because there is a need for some sort of a safety factor between the ultimate tensile strength and the design tensile strength. That is, this is the sort of stress that can be applied safely to the material while in use. The third equation that is presented on this line is the factor that I believe is standard for calculations in the USA. The ideal gas law is used, and this is modified for the compressibility, so PV over nRT is not 1 but Z. These other things that are presented here are then readily calculated. It is also possible to obtain a result for the third dot point on the slide, which is the mass percentage, including the mass of the cylinder (that is to say, the torpedo or whatever you want to call this tank) as well as the volume density, where the volume of the cylinder becomes an important consideration. So we are looking for something up in the top right of this diagram, where you see our targets. If you take the sort of advanced gas cylinder you find in laboratories that have hydrogen cylinders in them , they are aluminium, not steel, which embrittles. So aluminium 5083 is a material approved for building pressure vessels in Australia. This is its approved design tensile strength. Now as can be seen here in this diagram, at a level of around 200 bar is where the strength peaks. The dots here represent 100 bar, 140 bar, 200 bar, and then the performance decreases. The reason is that the structure gets too thick and heavy. This is, of course, about where gas cylinders go to: the 140 to 200 bar range. So that’s nowhere near the conditions required. Here is the state-of-the-art tank, presented according to my guesstimate from the announced performance of the tank which is at 700 bar. It is pretty well optimised: the highest point is 1,000 bar and you could design the tank to function at this point, but you would be losing mass density and it is also not really getting very close to the 2005 target. Owing to the performance of the material, it is possible to work out what the design tensile strength and density must be. It is hard to get this wrong. For example, the density indicated by the pointer on the diagram is close to the density of pure carbon fibre. This is a known factor that tanks are based on carbon fibre for strength, with other materials present in a composite form to actually seal the tank and keep the hydrogen inside. Thus this must be the design tensile strength they use. Let’s now say that you take pure carbon fibre – nobody has done this yet, and of course the gas would go straight through it. However let’s suppose you could make a material that was as strong as pure carbon fibre. This would be the predicted performance, and may even be close to the target set to be achieved by 2010. Though it would have to be run at 1,500 bar which is not a very palatable thought. If you aim to achieve the 2015 target you will get into rather esoteric territory. You will want a material whose ultimate tensile strength is nearly 10 gigapascals. Now, any mechanical engineers in the audience will know that the only materials that are around (and that one could contemplate) are things like carbon nanotubes, which have 130 gigapascals ultimate tensile strength – but of course there is the small matter of how one could make a continuous material out of these nanotubes in order to seal in the hydrogen, which is a mobile species. Then this would have to be run at about 5,000 bar. If you were to actually use carbon nanotubes themselves, all your problems would go away. You would have an essentially infinitely strong material, but you would still have to have between 2,000 and 3,000 bar pressure. These things may perhaps happen, but there is still an outstanding intrinsic problem here. That is, even without the tank, just to get the pressure needed to reach the 2005 target, you'll need 624 bar and for 2010 you’ll need 892 bar and for 2015, 2.5 kilobar - well sorry but I don’t want that in my gas tank! I also suspect that most of you don’t either. Furthermore, there is a lot of energy involved in compressing this gas; and there is a lot of energy transfer that needs to take place when you actually connect such a refuelling system up to your vehicle. So this looks like a technology that can not be considered in the long term. This is only an interim technology. You have already heard about liquid hydrogen, so I won’t dwell on this topic. Shown here is an old photograph which actually shows one of the prototype refuelling stations, and in the box are some of the problems associated with such fuelling stations. The energy content of the fuel is, of course, very good; at least the liquid hydrogen is all liquid hydrogen and the tank doesn’t weigh very much. On the other hand, the system is perceived as complex; it doesn’t meet the 2015 volume density target; and there are other realities associated with running this complex thermodynamic machine to liquefy the hydrogen. To return to our comparison, we appear to be left with the liquid hydrocarbons. Well these are fine in energy terms, but they contain carbon and this is exactly the substance we are trying to move away from. So, this group of new materials, presented at the left-hand side of the box, have only really been thought about seriously in about the last five years. In fact, some of them were really not thought about at all before, say, five years ago. All the materials in this box are if you like condensed matter storage systems. So by a process of elimination, it may be concluded that these materials are the ones that our research needs to focus upon. The issues surrounding solid-state storage are not all ‘beer and skittles’, if that has not already become obvious. In the case of solid-state storage in which the hydrogen molecule is dissociated, entropy will be altered to some extent which in turn leads to a rather high enthalpy of hydrogen storage. This is typical for the moderately strongly-bound hydrogen in lanthanum nickel 5. Some of these other more stable materials (that were presented on the right-hand side of the diagram I just illustrated) have rather higher enthalpies, but this is considered average. So if you take a very modest, small car in the order of the size of a ‘Smartcar’ and crunch some numbers about how much hydrogen will be required and how much power might be needed and so on, and then work out the amount of heat that must be dumped during refuelling - then the figure comes out to be 50kW for 1 kg of hydrogen, if this is achieved in five minutes. The DoE target is, in fact, 4 kg hydrogen in 2½ minutes, and that’s half a megawatt of heat. Perhaps that is ‘just’ an engineering problem, but it is nevertheless a real problem. At this point, I should also mention that gas tanks also involve large amounts of energy transfer, and they are by no means free of this kind of problem. Indeed these systems obey the first law of thermodynamics. So in this context, if there is adiabatic compression of gas (and adiabatic means that there is no heat allowed into or out of the system), and if this happens fast - as compression and combustion in cars happens fast, then the temperature has to change. So there are heat removal problems associated with these processes as well. You have heard about alternatives in which you store hydrogen not by absorption but by adsorption. Physisorption, and not chemisorption, if you like, is another way of referring to these alternatives. In these cases there is no dissociation involved, and therefore there is no enthalpy problem, but there is the problem of requiring a low temperature because of the weak bonding of the hydrogen molecule when it is physisorbed onto a surface. Here are some examples of ‘adsorbers’. The first are carbon nanotubes and although these are very controversial, there is still some hope for this area. After some very wild claims in a few very high-profile publications a few years ago, some people’s reputations began to oscillate on either side of the ‘border of respectability’. Yes, you can make and break your reputation in this field very easily by making a bad measurement. Indeed in this field, researchers are now working with a consensus range of a few wt% in carbon. So this is an interesting development, although nano-structured carbons are perhaps not very promising in relation to the goals set for 2015. A second type of adsorber includes metal-organic frameworks (MOFS) and similar materials which have a large surface area that is associated with links in a built skeleton. For these materials, it is not the volume that is important, it is the amount of skeleton. That is where the surface area is determined. Now depending on who you ask, anything up to 4 wt% may be achieved. I hadn’t actually heard about the 50 bar experiment; and I thank Cameron Kepert for pointing this example out. I wanted to do this experiment about three years ago although I was unable to obtain the necessary samples. So at present, there is some optimism about the use of higher pressures, though these are still very modest pressures as compared with pressurised gas tanks. In relation to isobars for hydrogen desorption from adsorbers, there is a lot of potential for improvement in this area. For example, a shift up the temperature scale can be observed with changes in the material science of these substances. Thus the important factor in these systems is to try and raise the temperature at which these materials will store hydrogen. So none of these materials really get very close to the targets, with the exception of the higher pressure MOFs (which we have just heard about) and the operation of these, are yet to be fully confirmed. Lithium-based systems (which I will have to discuss rather quickly) are however, very promising indeed. With these systems we are starting to talk about amounts of hydrogen that, in principle, are getting rather close to the 2015 targets. Hydrogen absorption and desorption by LiN are based on reactions like the one that can be seen at the bottom right of this slide. The problem with a lot of these materials is there are components, reactants or products which are stable, and so it is hard to access the hydrogen that is bound up in these substances. This is certainly a problem which requires tackling. Another type of approach that can be explored is to start to mix up some of these less conventional materials, like the amide shown here, with something more traditional like magnesium hydride. In pressure-composition isotherms, the reaction involved is effectively ‘destabilised’ (and this is a dangerous word). In other words, you move the pressure up to some more convenient value located at a lower temperature. So these interactions then form part of the game that we have to play and it may be noticed that the wt% is still not great. Other Li-based systems that are based on nitrides again are another interesting case because a mixture of some nitrides and a large amount of hydrogen are involved. It is also necessary to manipulate reactions with catalysts in order to lower the desorption temperature. This area of investigation is again promising. Now in terms of other lithium-based systems, lithium borohydride is a very stable material that is hard to decompose, indeed, you have to melt it. It is also hard to recompose. I was actually with Andreas Züttel in Fribourg, I think it was a few days ago (although I may be a little confused, having got back from Europe last night) and the progress being made there, in fact uses a ground glass catalyst to enhance this process. So, turning to the subject of the desorbed amount of hydrogen that occurs as a function of temperature for various heating rates - again we find that the ‘name of the game’ is to try to move the desorption peaks to a lower temperature. This is a very rate dependent process, as you will notice, and a complicated reaction mechanism is involved. Some of the most exciting reactions include mixing lithium borohydride with lithium amide. This is because large amounts of hydrogen are involved in these processes. However, another method that is even more promising involves the second reaction presented on this slide, where again large amounts of hydrogen are involved. In terms of the reaction involving hydrogenation and dehydrogenation of milled LiH, the process seems to be reversible. So really you can consider a sort of ‘melting pot’ in which it is possible to take any of the interesting materials that are shown here, stick them in the pot, add some catalyst, stir, heat – and then the products that emerge may possibly include something of interest.
Now on the subject of the National Hydrogen Materials Alliance, I will skip over these details. It may interest you that Australia is a major lithium producer. I come now to the conclusions. As far as I am concerned, these are that pressurised gas can really only be an interim technology. However, contrary to the rather depressing position that existed about five years ago (when we were still rather focused on metals), these less metallic metal-hydrides are indeed looking very promising. Thank you for your attention. Discussion Sukhvinder Badwal – Evan, if you take a 70 litre tank and pressurise it at 700 bar, that will give you roughly four kilograms of hydrogen. Certainly, Daimler are using this in their V Class car. It provides you with over 400 to 450 kilometres driving distance. So what is the problem with gaseous storage at 700 bar? Why do we want to go to storage of hydrogen in solids? Evan Gray – Well, as I said earlier, it depends on whether you consider the sector to be driven by DoE targets. The system you mention doesn’t get anywhere near any of the current DoE targets. So there is a choice. On the one hand, the targets could be re-thought (recalling that these figures were arrived at by negotiation with auto-manufacturers). Certainly in this case, if these targets were to be thrown out the window, then there would be a viable source of energy. On the other hand, if these targets were kept, then we would be nowhere near reaching them. So the choice is, I suppose, up to auto-manufacturers, and specifically whether they want to stick with the targets or not. I have to say again: as I deal with gas pressures like that every day, I am not keen on having them in my car, thank you. Note added in proof: The amount of hydrogen contained in a 70-litre tank at 700 bar at 300 kelvin is not 4 kg, but 2.73 kg, owing to the high compressibility of hydrogen (1.449 in this example), so the premise of the question is false. Michael Dopita – I am a little worried when you show me the actual layout of the hydrogen-powered buses in Europe. Having just recently been in London when there was a bus being blown up by terrorists, I wonder what would be the explosive effect of a similar attack on a hydrogen-powered bus of that kind. Evan Gray – I’m not the right person to answer the question. [Comment from the floor] – No more than petrol, Michael! Evan Gray – The stored energy is obviously high. The pressure in those tanks is 350 bar. Extensive destructive mechanical testing has been done by firing bullets into those tanks and so on, and because they are not made of brittle materials like metals that undergo rapid martensitic transformations or whatever – they are very different materials, they fail benignly – but yes, I wonder what indeed the effect would be of blowing one up. But I honestly cannot answer the question. Cameron Kepert – Evan, in combining these materials, as you say, in the mixing pot, does one end up with an average property between the two materials that you are adding together, or is it more complicated, in that you might be adding defects and instabilities into the structure and therefore getting better behaviours? Evan Gray – I don’t think anyone knows the answer to that question yet, because it is too new. Those results are really from the past couple of years, published in 2004, 2005, and 2006. In some measure you get an average result, but at the minute the important result is, if you like, the destabilisation of what started out as being very stable materials so that the hydrogen is too tightly bound and the pressure at which the hydrogen is liberated is too low. So these reactions have desirable effects on those. Some of the materials produced, by the way, are unknown. I mean unknown in both senses: people don’t know what they really are, but they haven’t been known before. Defects are always important in any other hydrogen storage material, so why should they not be important here? Harry Watson – I just wanted to comment on the effect of the destruction of these high-pressure tanks. Some well-documented processes are obviously taking place around the world to ensure that they are safe, and they are really quite undramatic. The tank inertia is so low compared to the metal objects coming out of a bursting petrol tank that there is a lot of comfort from that standpoint. But I suspect the eardrums are rather subject to unsavoury processes. I think they are at least as safe as an LPG tank, and I would say that as the owner of an LPG-dedicated car, I would feel equally comfortable. Les Field – Please join with me in thanking Evan for a very interesting talk.
Other speakers
Dr John Wright
Dr George Crabtree
Professor Cameron Kepert
Dr Sukhvinder Badwal
Professor Andrew Dicks
Dr Ben Hankamer Dr Catherine Grégoire Padró
Professor David Trimm
Dr Wes Stein
Professor Harry Watson |
© Australian Academy of Science