U.S. Department of Energy - Energy Efficiency and Renewable Energy

Fuel Cell Technologies Office

Hydrogen Storage Materials Requirements (Text Version)

Below is the text version of the webinar titled "Hydrogen Storage Materials Requirements," originally presented on June 25, 2013. In addition to this text version of the audio, you can access the presentation slides and a recording of the webinar (WMV 141 MB).

Alli Aman:
Thanks so much for joining today's webinar. I'm going to go through a few housekeeping items before I turn it over to today's presenters. First of all, today's webinar is being recorded, so along with slides and that recording, we'll post that to our website in about ten days, so definitely check back. However, I will also be sending an email out to everyone that has registered for this webinar when those do post to our website.

Everyone is on mute, so please submit questions throughout the webinar as they come up. We do have multiple speakers today, so if you have a particular question for a speaker, please indicate who that question is for, and we will cover those at the end of the presentation.

Today's presentation will go past the 1:00 hour, so if you have to log off at 1:00, not a big deal. I encourage you to submit your questions, because we still could answer those questions, and then you can listen to the recording once those post to the website.

And also, I encourage everyone—we do have a monthly newsletter, and if you're not signed up for that, I encourage you to visit our website and register to get those monthly newsletters, and that will also keep you in the loop of future webinars.

So on that note, I'm going to turn it over to Ned Stetson. He is the hydrogen storage team lead for the U.S. Department of Energy Fuel Cell Technologies Office in Washington, D.C. Ned?  

Ned Stetson:
Thank you, Alli. I would like to welcome everyone to this webinar on the hydrogen storage materials requirements to meet the DOE hydrogen storage targets for onboard light duty vehicles. As Alli said, I'm Ned Stetson, team lead for the Hydrogen Storage Program at the DOE.

For a number of years, the DOE has carried out a very comprehensive program of developing advanced hydrogen storage technologies to enable hydrogen fuel cell vehicles to have performance on par with that of conventional gas tank fueled vehicles. A key challenge has been, and still needs to be, having a cost effective hydrogen storage technology that's compact enough to provide at least a 300 mile driving range without compromising the passenger and cargo space, as well as meeting all the other performance criteria.

Between 2005 and 2010, a significant portion of the DOE's hydrogen storage development efforts were devoted to developing advanced hydrogen storage materials, with the research being carried out in three materials Centers of Excellence, one focused on each of the material classes: hydrogen adsorbents, chemical hydrogen storage materials, and reversible metal hydrides.

In 2009, the DOE established a fourth Center of Excellence to investigate the engineering aspects for incorporating materials into a complete and viable system. So while today no material-based system has been identified as able to meet all of the onboard storage targets, the efforts of the Hydrogen Storage Engineering Center of Excellence have led to a much better understanding of the material level properties that are needed in order for a complete system to meet the challenging onboard vehicle targets.

For reversible metal hydrides, the Engineering Center of Excellence partners have previously presented some results on the material requirements through various conference presentations and general publications. So today, we've invited several members of the center to present on the properties needed for hydrogen adsorbents and general chemical hydrogen storage materials, so the complete system can meet the onboard targets.

As Alli mentioned, you can submit questions during the presentation, and at the end we will try to answer as many of them as time allows. When submitting questions, please indicate which speaker you'd like them to be directed to. We do ask that you limit your questions today to the presentation, as we will not be responding to any questions on the planned hydrogen storage funding opportunity announcement.

So today we have speaking Dr. Don Anton, who is a senior advisor—advisory scientist at Savannah River National Laboratory, and is the director of the Hydrogen Storage Engineering Center of Excellence. We have Dr. Troy Semelsberger, who is a chemical engineer at Los Alamos National Laboratory, and he is the lead coordinator or system architect for the chemical hydrogen storage system efforts within the Center of Excellence.

We have Mr. Kriston Brooks, who is a senior engineer at Pacific Northwest National Laboratory, and he is a main modeler who developed the transient models used to evaluate the chemical hydrogen storage system within the Engineering Center of Excellence. We have Dr. Don Siegel, who is an assistant professor in the mechanical engineering department at the University of Michigan, and a lead coordinator or system architect for the adsorbent system efforts within the Engineering Center of Excellence. Prior to joining the University of Michigan, Don led the fuel cell and hydrogen storage materials group at Ford Motor Company.

And then finally, we have Dr. Bruce Hardy, who is an engineer at Savannah River National Laboratory, and a main modeler for the hydrogen adsorbents, and leading the transport phenomenon technology area within the Center of Excellence. So with that, we have a lot to do today, and I'll turn it over to Don Anton.

Don Anton:
Okay. Thank you, Ned. Next, please.

[Next slide]

I just wanted to give the—reiterate what Ned has said, that this work is fully funded by the Department of Energy through the Office of Energy Efficiency and Renewable Energy and the Fuel Cell Technologies Office. And what we're here today is to give guidance to the materials development community, all of those listening out there, as to what the important materials characteristics need to be to meet the 2017 DOE technical targets for onboard hydrogen storage for both the adsorbent materials and the chemical hydride materials. Next.

[Next slide]

The Center of Excellence has used the current materials data as we understand them and applied the thermal management modeling, looked at the balance of plant for the system which needs to go around that to put it into a useful form for the automobile to use, and then used models to model the fuel cell and the vehicle performance. And now we are feeding all of that information back to the materials engineering community so that they can develop better materials.

I'd also like to just acknowledge all of the work from all of the centers, the participants. It's been a very open forum, and people from all of the other groups need to be recognized for their input into this program. Next.

[Next slide]

One has to remember that we have the technology targets, and I'm not going to go through those in any detail at all, but those technology targets are for the full system. So the volume of the system is composed of the sum of the media and the components, and we have been looking at the components of the system, the mass, the volumes, and our structure as to how these would look.

One has to remember that if one had a significantly different system configuration, that that would affect very much the media requirements to meet that system. So we're going to be discussing first the system that we have, so the systems that we've been looking at, and then the materials that will be necessary to meet those system requirements. Next.

[Next slide]

So the general outline is for each of the groups to first define the systems, define the technical barriers that have been identified for those systems, and then finally to identify the materials properties that will meet the target based on those systems defined. Ned has already introduced the speakers, so I won't go through that again.

[Next slide]

And with that, I will pass it off to Troy Semelsberger.

Troy Semelsberger:
Good afternoon. I'm assuming everyone can hear me appropriately. I'll present the chemical hydrogen storage material requirements, as Ned and Don alluded to, for our system, and those will be discussed in detail. Next, please.

[Next slide]

Just to mention that Kriston Brooks, who's performed a lot of the modeling work for the transient system modeling, employing U.S. drive cycles, contributed to this as well. There'll only be one presenter, and I encourage any questions to be addressed to either Kriston or myself. So the key takeaways for today are really the parameters that will allow the materials to meet our system design assumptions, operating conditions, and expected components for our system. So the biggest parameter that most materials researchers are going to cling to are most likely the material capacities, and there are three flavors of that I'll present. And the second is most likely the kinetics. What are the kinetics associated with those materials to allow the hydrogen performance flow rates that are required with the DOE performance targets?

And then the one that I won't talk about but is included in the table is regeneration efficiency. The regeneration efficiency of the off-board materials should be at minimum 66.6%. So that's a number that will be viewed upon as being critical for viable chemical hydrogen storage systems to be commercialized. Next.

[Next slide]

So the primary objective, again, is just to provide guidance to the materials researchers with a number of different parameters that will allow them to meet the DOE performance targets. Our approach is a four step process where first it was to develop the integrated storage system with all the necessary components, and then wrapped around that is the system model developed by Kriston that will allow us to predict the system performance using various drive cycles and conditions to assess the transient behavior of the automotive system.

The third step, actually, we identify and size the components that are material independent, and those being, for example, the reactor or heat exchangers. And what this does is actually removes the component from being material specific. In most cases, we were dealing with ammonia borane and alane, but all of those components were specific to those two materials. The third step allows us to remove us from any material specific hydrogen storage material to make it more general that could be applied to the general class of chemical hydrogen storage materials. And with that sizing, we can actually back out material properties to fit that component's size, mass, and performance.

And lastly, the—given the component masses and volumes, we can actually back out what the material capacity should be to meet the volumetric and gravimetric targets of DOE.

[Next slide]

So next.

[Next slide]

So I'll first talk about the storage system, which is—includes the primary system components along with all of the balance of plant components, balance of plant components being valves, pumps, sensors, etcetera. The system components of primary interest that are material dependent include the volume displacement tank up in the upper left, where the cursor is, or just was. In the bottom middle is the reactor, followed by the phase separator. Then we have our gas and liquid radiators, ballast tank, and hydrogen purification system.

Now this system is expected to be a system that's general enough to apply to all chemical hydrogen storage materials. They may fluctuate in mass and volume, but these are the essential ingredients of our system. Next, please.

[Next slide]

So here's an itemized bill of materials list for our systems. I'm not going to go through it, but we have identified manufacturers, vendors for each of the components that we deem necessary for a viable system. Next.

[Next slide]

So you can actually bin these components into three categories: material independent components, which we refer to as the BOP. Those include the valve sensors. The material dependent components, which are the reactor, purification system, ballast tanks, heat exchangers. And then the last is the system independent material properties or components, and those include the material storage capacity, the regeneration efficiency, fuel cost, and shelf life.

Now on the table to the right, you can see that there are two systems identified: the baseline system, which was shown earlier, and then our idealized system. So the only things that change between the baseline and idealized system are shown in italicized black elements, or the entries, and so for the idealized system, the purification system is zero mass, zero volume, which indicates that—the assumption that the material does not produce any impurities that require scrubbing.

The second change is a reduction in mass and volume from the reactor. It would be actually cut in half. And you'll see the impacts on the material capacities required with that reduction in mass and volume. So next.

[Next slide]

Now if you were to put these in pie chart form, this is what it would look like. For the baseline system, you can see on the left is the volume pie chart, where the media takes roughly 50% of it, and then of course your reactor purification system, ballast tank, and BOP rounding out the rest. Now if you can see the exploded black wedge, particular attention should be paid to that, because we have an unused volume of 33.4 liters. That means, you know, for lack of a better description, you have a bigger trunk space, right? I mean, or you're sitting in first class. So that volume is available for other amenities for the vehicle.

Now on the pie chart on the right, which is the mass, we actually fill out all of the available mass within the DOE targets, which is 102 kilograms. So as you can see from this plot, our primary driver in meeting the DOE targets is really the mass. We have a surplus of volume, but our mass is really constrained, and that is our primary driver. So next.

[Next slide]

So following on, if you do the same process for an idealized system, you can see we've removed the purification system, and we've also reduced the reactor mass and volume by half. So you can see now that the media mass is taking more volume and more mass of the entire system, and in doing that, we can actually reduce our material capacity by that mass and volume that we gained from reducing the reactor and the purification system. And that'll become evident as we proceed.

[Next slide]

So next.

[Next slide]

So going onto the material properties, after we have our systems defined, so when it all boils down to it, you have for an idealized system your system mass being 30.6 kilograms—again, that's with no purification and a reduced reactor mass. Our baseline system is 36.3, and those are actually indicated, if you look at the assumptions, indicated in the parentheses. So again, for the idealized system, our fixed reactor mass is 2.5 kilograms. There's no purification mass. So that total mass, including the BOP, results in 30.6 kilograms. And again, a primary assumption is that we have no phase change in the media, which would be catastrophic to the system.

So if you turn your attention to the plot in the upper right, so again, our primary drivers are mass. So if you look at 30.6 kilograms on the ordinate, which is the green dashed line, go over to the DOE 2017 target envelope, and then drop vertically down, you will get a material capacity that's required, a minimum required capacity of 7.8% or 0.078 grams of hydrogen per gram of liquid.

Now if we increase that mass to our baseline system, which equates to 36.3 kilograms, that is represented by the solid green line. With that increase in mass of our system, it actually causes the material capacity to increase in order to meet our DOE 2017 targets, which is 8.5 weight percent, or 0.085 grams of hydrogen per gram of liquid.

So you can see the sensitivity in the system mass on what the net usable hydrogen of the weight—of the liquid is required to meet our DOE 2017 targets. So there are other ways to make a fluid hydrogen storage material, which is presented on the next slide, and that is being a slurry.

[Next slide]

So if you have a material that you want to put in slurry form, the material capacity of that solid material has to increase, and there's only so much slurry that—mass fraction that you can have in order to accommodate the gravimetric capacity required to meet the DOE 2017 targets. And that's actually bounded by the slurry volume fraction, which is the plot on the lower left, where you really can't have a volume of your solid exceeding that of your liquid. I mean, the rheological properties associated with that, you know, deviate too greatly, and most likely represent a solid movement on board the vehicle.

So if we actually fix the maximum slurry volume fraction to 0.5, that has to correlate to a slurry mass fraction, and we've created two bounds, an upper and a lower bound, that they're represented by a high gravimetric capacity of your material, and a low density of your carrier—that is the upper bound. Then if you have a low density material—this is the hydrogen storage material—and a high density of carrier, that creates your lower bound.

And by doing that, you fix the slurry mass fraction, which is, you know, I've identified at 0.35 and 0.7. Now if you translate that over to the upper right, your range on your abscissa is 0.35 to 0.7, so that's our likely slurry mass fraction. And the material capacity associated to meet our 2017 targets for idealized and our base system are shown. The dashed line is for the idealized system, and the base system is shown in red.

So an assumption that we have to make is the slurry is homogeneous and non-settling, which is an ideal slurry. So with those values, we can actually start backing out what the net usable weight fraction of hydrogen for that solid material has to be, given those constraints. And it turns out that it's about 11.2%, or 0.112 grams of hydrogen per gram of solid, for the idealized system. If we were to apply this to our baseline system, which is 36.3 kilograms, it actually bumps that material capacity up to 12.1%.

So those are the ranges that are expected for a slurry if they were to be viable options for an onboard chemical hydrogen storage system. So next.

[Next slide]

Finally, you can actually produce a solution. The plot is slightly different on the abscissa, where the range extends to 0.8, and that is the maximum solute mass fraction that we expect for a dissolution of a chemical hydrogen storage media. And it should be pointed out that both in the slurry case and in the solvent case, the carriers or solvent are non-hydrogen bearing. So if that were to be the case, this—these would change. You could actually decrease the amount of your gravimetric capacity of your solute or your solid material for slurries. So again, no phase change or—the system mass, again, is for the idealized and the baseline system, represented by the dashed blue line and the solid red line.

So again, going through the same exercise, if your solute mass fraction has to be between 0.35 and 0.8 gram solute per gram of solution, that correlates to a net usable weight fraction of your solute to be a minimum of around 9.8% for the idealized case, and 10.6 for the baseline. Next.

[Next slide]

So shifting from the material capacities from a system independent viewpoint, we now have to look at, given our sizing of our reactor and our other components, what are those parameters—what parameters are necessary to meet those sizing requirements? So for the reactor, we assumed a shelf life of greater than 60 days, the temperature 60 degrees C, and a max conversion of 7.2%. The volume of the reactor at 175 degrees C and a conversion of 99% had to be less than or equal to four liters, and that reactor has to produce a maximum flow rate of 0.4 moles of hydrogen per second.

Our variables were the activation energy, pre-exponential factor, and reaction order, so what are the kinetics that we need in order to satisfy the constraints of shelf life and the PFR volume? And in doing those, you can generate Arrhenius plots and then back out what the activation energy and pre-exponential factor are required to meet the 2017 targets.

And in doing that, you actually come out with a range of 28 to 36 kcals per mole of hydrogen for the activation energy. The pre-exponential factor is four times ten to the ninth up to one times ten to the sixteenth, and these apply for power rate law reaction orders between zero and one.

Now you could actually have a higher activation energy, and if you were to apply a catalytic process, you could lower that into this range. So—but those aren't taken into account, because adding that variable into this, it's ill-defined and cannot be quantified. So the final activation energy and pre-exponential factors for the ranges are shown. Next.

[Next slide]

So the exothermic heat of reaction is a variable that we're fixing because we do not want the material to exceed what we've put in as a default, as 250 degrees C. And this system is actually bounded by the ammonia borane system that has a high heat of reaction as well. So the maximum system temperature was maintained at 250 degrees C, with an inlet temperature of the material at 24. And we also assumed a maximum recycle ratio of 50%, and this allows you to actually cool your material.

So in looking at the heat of reaction to satisfy these constraints as a function of material heat capacity, material hydrogen capacity, and heat capacity, you can see the surface plot. And for all of these heats of reaction shown, you know, you can accommodate our constraints of a maximum system temperature of 250 degrees C. So we can actually conclude that if your heat of reaction is greater than minus 27 kilojoules per mole of hydrogen, you will not meet this constraint, so anything less than that, we're okay. So next.

[Next slide]

Following along the same lines is the endothermic heat of reaction, and in producing the property range for this parameter, we actually resorted to looking at the onboard efficiency. And again, the two plots on the right, one is a 2.5 kilogram stainless steel reactor, which is the idealized case. The 5 kilogram stainless steel reactor is for our baseline case.

So the assumptions in order to do this were many. Here are a few.  We did not assume any heat recovery. We have the fixed reactor masses. We actually identified a cold startup, and that is defined as the temperature of the reactor, which we set at 175, and then minus the T ambient, which is 25 degrees C, which results in a delta T of 150 degrees C.

And we actually set the number of startups per day to be four, and that's the number of times you actually start your vehicle. So if you start your vehicle to come to work, you start it up to go to lunch, start it up to come back, and then start it up to go home, that's four. So the average miles driven per day is 41. And again, we assumed a neat liquid with a heat capacity of 1.6 joules per gram Kelvin.

Now if you do this and you plot the onboard efficiency as a function of heat of reaction and startups per day, you get the two plots. For 2.5 kilograms reactor at four startups per day, you're looking at about 17 kilojoules per mole endothermic heat of reaction, while still maintaining an onboard efficiency of 90%. Now you can see if you had eight startups per day, you're looking just over 14, maybe 14.5, 15 kilojoules per mole of hydrogen.

Now if we assume our baseline system that includes a 5 kilogram stainless steel reactor, again, if you assume four startups per day varied with the heat of reaction and onboard efficiency, so that actually drops because of the mass that you're required to heat the reactor, the additional mass of the reactor, which is 2.5 additional kilograms, it actually turns out that you need about 15 kilojoules per mole of hydrogen endothermic heat of reaction in order to maintain an onboard efficiency of 90%.

So if you look at the exothermic and endothermic heats of reaction, there is actually a wide range that can be accommodated while still meeting the onboard efficiency or outlet temperatures of the reactor. So next.

[Next slide]

This slide actually just demonstrates the sensitivity of the volume displacement tank as a function of the final composition density and the final composition with respect to hydrogen capacity. So we actually fixed our volume displacement tank to be 6.2 kilograms, and that was actually a fallout—was born out of the ammonia borane. If you have about 8 weight percent media hydrogen capacity, and the media density is slightly less than 1, you follow the abscissa and ordinate on that and then look at the tank mass on the Z coordinate, it's about 6.2.

So anything heavier or anything with higher gravimetric capacity will certainly meet the—our fixed tank mass of 6.2 kilograms, and that actually correlates to a hydrogen density of the media greater than 0.07 kilograms of hydrogen per liter. So next slide.

[Next slide]

Rounding out the system components is the purification, and we actually fixed for the baseline system a fixed mass of 3.2 kilograms. And we did assume that it's an adsorbent-based technology, and the purity being 99.97%. And in this case, we actually had a replacement frequency of 1,800 miles, meaning that after 1,800 miles you have to replace that, not unlike an oil filter.

So in doing that, the deeper you dig with respect to trying to quantify the maximum impurity concentration, it turns out that it's actually too difficult, and there are too many assumptions involved that will allow anything to be reasonably acceptable for a range, and that's actually shown in the surface plot, where depending on the capacity of—in this case adsorbents—it is a function of your molecular weight, because you've got to keep in mind that everything is mass dependent.

So if you have a really heavy impurity that's produced, say 90 grams per mole, you know, and if you assume a 50—a 0.5 capacity, you know, you're looking at about 1,000 ppm. But on the other hand, if you assume the same capacity, but if it's at 40, you know, you're at about 2,500.

So there's too many variables in order to quantify what the impurity level should be, and not to mention that the impurity concentration for a fixed mass is also a function of what type of impurity. Is it a fuel cell impurity or is an inert diluent? Does it require scrubbing? The degree of scrubbing, for example, if it's nitrogen, would prove very difficult compared to something say like ammonia or borazine. So the chemical and physical properties of the impurity are going to strongly dictate not only what technology, the capacity, and the coefficient of performance of that adsorbent base or purification technology.

In addition, does that impurity require to be recycled, to put back into this closed loop recycle for the chemical hydrogen storage material? What is the cost? So all of these kind of play a role in determining what impurity concentration is allowable, and I think for the most part they'll be actually dictated on a case-to-case basis. I was hoping to actually get a range out there, but like I say, it's just—it's too ill-defined in order to accurately capture that impurity concentration. So next, please.

[Next slide]

So in summary, here are all the material properties that we feel are important that I think the materials researchers can use and leverage in their research efforts. I've listed all the assumptions and the influences on the components. And if there are any other questions I think with respect to these, please feel free to obviously give us a call.

The most important I think is material capacities, which are the primary driver right now, because of the mass implications of the material capacity, so—not to mention that the kinetics will also play a role. If it sits there and doesn't do anything, where it will eliminate the potential to meet our minimum hydrogen flow rate, we'd require a different system. So—and I'd mention all these material properties are based on our system, our system assumptions, and the conditions that were presented in the context of this presentation. So next.

[Next slide]

So the next step is hopefully the materials community can leverage these material properties in their development of new materials, and you can evaluate your novel materials relative to what's presented. But also, as the materials show promise, I urge everyone to seek Kriston or myself out and allow the materials to be put into a higher fidelity approach, and that being a system model developed within the Engineering Center of Excellence. And actually, that may provide additional guidance on certain parameters that aren't really shown right here today. So next slide.

[Next slide]

So please read the disclaimer, and with that, I'd like to thank Ned and EERE and the Fuel Cell Technologies Office for their continued support, and that concludes my presentation.

[Next slide]

Thank you.

[Next slide]

Alli Aman:
And Don, on that note, we'll turn it over to Don Siegel.

Don Siegel:
Okay. I hope everyone can hear me.

Alli Aman:
Thank you. We can hear you.

Don Siegel:
Great. Okay. So I'm going to provide a brief overview of some of the work taking place in the engineering center with respect to adsorption-based hydrogen storage systems, and the goal of the work is really just to provide some context for what Bruce Hardy is going to talk about in the next portion of this talk related to materials operating requirements. So what Bruce will say is informed by about four years of engineering analysis focused on adsorption-based systems. Next slide, please.

[Next slide]

So the adsorption system is, you know, being developed in parallel to the chemical hydride system that Troy talked about, and the goals are roughly the same. What we're trying to do is model, design, and construct an adsorbent-based hydrogen storage system that has the potential to meet the DOE 2017 targets. And I say potential to meet because, as I'll show you later, you know, we're trying to do a very good job with the engineering, but ultimately, how far we can get is going to be impacted by the materials performance, and that's one of the reasons why we're having this webinar today, is to show that some more work is definitely still needed, and to hopefully guide materials developers on the right path to solving those challenges.

[Next slide]

So in addition to that, we've been trying to quantify and understand design tradeoffs. So for example, there's some tradeoffs we've found between gravimetric and volumetric capacity. I think we need to go back two slides, still.

[Next slide]

There we go. And another example would be capacity and cost versus filling time. If you want to fill your system very quickly, maybe you need an elaborate heat exchanger which takes up space, has mass, and could be expensive. What we'll talk about today, like I've already said, is using the engineering analysis we've developed to identify those materials properties that are most strongly related to system performance. Okay.

[Next slide]

So you sit down and you decide you're going to design an adsorption-based hydrogen storage system. You have a clean sheet of paper, and you soon realize there are many challenges or many decisions you have to make. Some of those are summarized on this slide. So starting with the material, there's a question of, well, what form will that material have? Is it going to be a powder? Is it going to be a pellet? Is it going to be some sort of monolith, maybe like a hockey puck or a cheesecake? So that goes inside the tank.

Another question is what are the operating conditions, and how will they influence your choice of tank material? So one can have an expensive, lightweight carbon fiber tank, which would allow higher pressures. On the other hand, maybe something which is an aluminum type one tank, or just a metallic type one tank, could save some dollars, but would be heavier. Understanding that tradeoff is of course quite important.

And then finally, what goes inside the tank in support of the material, so the heat exchanger. So for the adsorbents of today, we don't have very high thermal loads, but they're not zero, and so we do have to think about what will be the design, how efficient could different heat exchanger designs be? These could range from something rather simple, for example, a resistance heater, to a more elaborate design, which I'll talk about later, the so-called MATI system. Okay.

[Next slide]

There's also the question of what composition of material should be used. And so this was one of the first choices that the center made, and to make this decision, we weighed a number of different factors. For example, what could our industrial partner, BASF, who is going to be responsible for synthesizing the adsorbent, what were they comfortable with? What could they make? What sort of materials were available, and what were known about them?

So at that time, there were a handful of metal organic frameworks which were known and had been moderately characterized. We also knew about quite a bit of work on—relating to activated carbons. And at the end of the day, the decision came down to a decision between MOF-5 and activated carbons, and we decided to go with MOF-5 because, well, it's a very well-studied material. Its performance is a bit higher than activated carbons. There was also some opportunities maybe to explore, or maybe MOFs offer more of a wider phase base for future development, maybe a wider variation in composition and structure could be possible.

And so at that time, we tried to make some assessment that would of course choose a MOF that would give us a nice balance between gravimetric and volumetric. At that point in time, we didn't have a very complete analysis of the different MOF possibilities. And what this slide in front of you shows now is a more extensive analysis, which shows that MOF-5, which you can see in the center of the plots, near the maximum in volumetric density, is actually a reasonable compromise between gravimetric and volumetric performance, although there are a few other compounds which we recently identified that could perform a little bit better. And these of course could be possibilities for future developments. So next slide.

[Next slide]

So with—moving forward with MOF-5, we've been focusing now on specific designs for the adsorbent system, and essentially, what we've tried to do is quantify how different designs would impact the ultimate performance. And so David Tamburello with some help from Mike Veenstra have come up with a nice system ranking metric, which we've exercised to essentially down-select two different designs which we hope to pursue in the future.

One of them is shown in front of you here. This is the so-called MATI, or Modular Adsorption Tank Insert system. So the advantage of this system is it allows us to densify the MOF, to take MOF powder and press it into pellets or pucks, thereby increasing the volumetric capacity, and to cool and heat those pucks in a relatively efficient fashion. So let me show you a little bit more about the internals of this design on the next slide.

[Next slide]

So what the MATI does is basically consist of a series of metallic plates which are for either heat or cooling distribution, as well as hydrogen distribution. You see those blue plates in the middle of this slide. In between those would be sandwiched a puck of, say, activated carbon, or like I just mentioned, MOF-5 pucks, and those are stacked up and can fill space rather nicely.

So some of the initial design work you can see on the bottom left with the fabrication of the pucks, which are about five centimeters in diameter, and then some thermal testing has been done that's shown on the right. For example, at the top right, you can see a puck which has been cryogenically cycled, and then an assessment of the thermal properties of that system, comparing measured temperatures using thermocouple leads, which you see at the top right, to thermal models. Next slide.

[Next slide]

So another example of the system—of one of the systems we're pursuing, which is—basically bookends the design space and is quite different from the MATI system, is what we call the hex-cell and flow-through system concept. So here, instead of using a densified adsorbent, we're using an adsorbent powder, which would fill the hexagonal heat exchanger manifold you see at the bottom left. And in this case, we'd be cooling the system by flowing cold hydrogen around those MOF powder particles, so using a convective cooling approach, and then heating the system by using resistant heating—resistive heating rod, which would go down the center of the hexagonal manifold, but the heat would be conducted through that metallic system to bring about desorption.  So next slide.

[Next slide]

So with those two systems, we have ways of exploring different heating and cooling approaches, and different morphologies for the adsorbent, and to give you some sense of where we are right now with system performance, here I'm showing a spider chart, which compares the performance of the MATI system with respect to the DOE 2017 targets.

So you can see that there is some white space in this slide, in particular, some barriers, like gravimetric costs, volumetric, and loss of usable hydrogen still stand out. So we still need some development on the material side to help close these gaps. But again, we've tried to be as efficient as possible in engineering the system, and one thing that we've done, which is not well—which I haven't had time to describe is, say, lowering the cost by using a lower pressure system going down to say 100 bar using a type one tank. And that's in fact—if you can see the difference between the light blue and the blue on the system cost wedge, that's one of the advantages of the design change which I just mentioned. So next slide.

[Next slide]

Another way to visualize the challenges or the obstacles which still remain and why we still need further materials development is shown in this series of waterfall plots. So in each of these plots, we're comparing the DOE targets to the current status of our system or its projected status. So the red lines represent the targets, and the blue dots represent where we are now or where we were and where we would like to be.

So in particular, you can see, looking at the bottom two plots, we started off about two years ago at point A, quite well below the system gravimetric and volumetric targets. We've made some progress going from A to D, where D represents our current state of affairs. We've made some improvement there, while dramatically reducing the cost, if you see the top plot on the left.

So we still have a ways to go, and what this shows you is that to get to those DOE targets for capacity, you roughly need about a 60 to 80 percent improvement beyond the properties of current MOF-5. So next slide.

[Next slide]

And so what we're doing next in the engineering center as we move into phase three is to validate the system models we've—I've just talked about by building complete hydrogen storage systems around these two concepts, and testing the integration of them to make sure they perform in the way we hope.

[Next slide]

So with that, I think I will stop and pass it off to Bruce.

Bruce Hardy:
Okay. Let me get my presentation up here. I assume everybody can see this. And I would like to thank Don Siegel for introducing a number of concepts that I'm going to be talking about here. And also in this—with this work, I'd like to acknowledge the work by Claudio Corgnale and David Tamburello, who've made major contributions to this particular topic, the adsorbent acceptability envelope. The objective of this methodology is to identify a couple of adsorbent and storage vessel properties that make it possible to meet performance targets. The—you can certainly do this with detailed models, but—

Alli Aman:
Bruce, I hate to interrupt you, but we can't see your screen, so would you mind if I run—if I run—

[Crosstalk]

Bruce Hardy:
– not working yet. Okay.

[Next slide]

Well, we can go ahead and run whatever—yes. Go ahead and run—

Alli Aman:
Oh, okay. There it is.

[Crosstalk]

Ned Stetson:
It showed up.

Alli Aman:
It just popped up. Thank you.

Bruce Hardy:
I don't know what I did, but okay. Anyway, I'll just continue on, if that's okay.

Again, what the—the purpose of the acceptability envelope, the adsorbent acceptability envelope, is to identify material adsorbent and storage vessel properties that make it possible to meet performance targets, and then the idea with this is I should add quickly, because you could certainly do this with detailed models, but the problem with a detailed model is they take anywhere from several weeks to several months to develop, and it's very likely you might overlook some characteristics of the adsorbent or the coupled system that would make it impossible to ever meet the performance target.

So for efficiency, we have a system to identify rapidly the necessary properties of the adsorbent and storage vessel that would make it possible to hit the targets, and then we go on with detailed models. This is accomplished in two stages.

The first stage, we identify isotherms that tell you that the adsorbent can possibly hold enough usable hydrogen. By usable hydrogen, this is different than total. When we charge a tank, we start at a certain—we charge a tank, we go to a certain temperature and pressure. Then with the discharge, we go to a second temperature and pressure, a different state, if you will. And the amount of hydrogen released between that initial and then final state is the amount of hydrogen that we can of course send to the fuel system in the vehicle, to operate the vehicle.

It's not enough to just talk about total stored hydrogen, because you might go from the initial to the final state and still have a lot of hydrogen retained in the system. If you can't access that hydrogen stored by the adsorbent, if it's just held by the adsorbent as a high amount of hydrogen adsorbed between the initial and final state, you really can't use it. And again, this depends on initial and final states.

And we determined the amount of usable hydrogen through isotherm parameters, which means you fit a—fit your data to the isotherm parameters. And we talked—we speak to the isotherm parameters when we characterize what is a—what is an acceptable system, a system that meets the performance targets or not. And so far, we've considered UNILAN and Dubinin-Astakhov-Radushkevich isotherms. We could certainly consider others.

The—also, in addition to just determining isotherm parameters that would allow you to have a sufficient amount of usable hydrogen to meet the performance targets, we could also determine optimal parameters, parameters that tell you what could the best adsorbent look like. Another thing that these isotherms were used for in the actual calculation—sorry, in the scoping calculations, the second stage more than—certainly not in the first one, and also for the detailed calculations, is differential excess enthalpy of adsorption. That's actually what we used to perform the thermal analysis, and the—this quantity is determined directly from the isotherm, so you don't need to add an additional measurement feature into the model. You can get it right from the isotherm.

The second stage is to determine the coupling between the adsorbent and storage system that's required to meet the technical target. This requires everything in the first stage, plus you have to have some idea about how you want to design your system for charging and discharging in terms of your pressure vessel, the mass of the pressure vessels, the spacing between heat transfer components, and how you're going to operate it.

Due to time limitations, what I'm going to talk about in this presentation is just mostly stage one. We're going to give some results for stage one, and stage two I'll discuss in brief.

[Next slide]

Now in terms of optimization, what's the best type of adsorbent we could possibly have? What we do is we specify, again, the initial and final states by temperature and pressure. We determine the optimal parameters with respect to the amount of usable hydrogen—again, the amount of hydrogen delivered between the initial and final states. For UNILAN, which I'm going to talk about here, we optimized on—and in the succeeding slide—we optimized on the limiting adsorption, N max, and the energy, the adsorption energy, the maximum site energy for adsorption and minimum site energy for adsorption.

We can also optimize with respect to other parameters, like say pore volume and entropy change. And I mention the word constrained, because these are optimization processes, and when we optimize with respect to an objective function, say the usable amount of stored hydrogen, the mathematics doesn't care if it gives you something unrealistic, like a negative pressure. So you have to constrain this to ensure that you have physically realizable parameters.

And also, we can—we can optimize on the initial and final temperature states to give a better optimization or to give additional performance operating ranges for the system. Again, those have to be constrained. It might give you a very, very high pressure that you really physically can't deal with with a pressure hose.

The isosteric heat was calculated in the next slide, so for these optim—at the optimized parameters. So what the material developers would need to focus on is if you have an existing material, you'll fit your data to isotherms and see how they compare against optimal or acceptable isotherm parameters, or if you're going to develop adsorbents, this acceptability envelope can provide target parameters that you want to design to, to be able to meet the technical performance target.

[Next slide]

The UNILAN isotherm model, and I hope you can see my pointer here, but this is the UNILAN isotherm model. This is the absolute adsorption. N max is the limiting adsorption. You've got the gas constant times temperature at the maximum and minimum site energies. The enthalpy of adsorption, and that tends not to change a lot with the adsorbents that we know about, and that's via paper by Bhatia and Myers.

This is the pressure, total pressure, and this is the reference pressure. It's one bar. The total amount of stored hydrogen includes the amount of hydrogen held by the adsorbent, plus also the hydrogen that's held in the interstitial space. When you form a packed bed of adsorbent, there's space between the crystals of adsorbent. That volume can be a significant amount of stor—can give you a significant amount of stored hydrogen, and that's included in the total amount of hydrogen that's stored.

The usable amount of hydrogen is the difference between the total hydrogen stored at the charge state and the total hydrogen stored at the discharged state. Again, we—for this particular example, the charged state is at 80 bar—80 K and 60 bar; the discharged state's at 160 K and 5 bar. We constrained the limiting adsorption to between 0 and 120 moles of hydrogen per kilogram of adsorbent. That was just to put a bound on it to make sure that we have a reasonable range for this particular case.

The minimum energy, site energy is positive. The maximum site energy is the minimum site energy incremented by one. The reason that increment is there is because UNILAN isotherm results in a singularity in the isosteric heat if E max happens to be E min. Now you can change that increment to a much smaller value. You could go say six places left—right of the decimal point, and it makes very little difference in the results. So you just don't want E max to be exactly E min because of a mathematical artifact of the UNILAN isotherm.

Now in this table, what we see here is this is the—first row is the parameters for MOF-5, so the isotherm parameters for MOF-5, and the last column is the usable hydrogen in terms of kilograms of hydrogen stored per kilogram of adsorbent. At the optimal state, what happens is we ran up to the top of the range for N max, and E max and E min because essentially equal, as they—but as this value, if you go beyond those values, you start to drop your capacity. Delta S zero was held fixed, but again, that could be optimized again. And you can see that the usable hydrogen is increased by quite a bit.

This result is consistent with the paper by Bhatia and Myers, who didn't look at temperature, they only looked at pressure swing adsorption, and they found that there was optimization when E max was equal to E min, and essentially that means there's no heterogeneity for the adsorption sites. They all have the same associated energy of adsorption.

[Next slide]

Now again, our optimized UNILAN parameters are shown in this table. The common definition for the isosteric heat that people generally use—material developers key in on this a bit, and they use this definition. Myers doesn't like it because if you do heat transfer calculations, you need something different, but we'll—this is the definition that's in most common use.

And what you notice is this isosteric heat, when you do this calculation, the isosteric heat is constant with respect to pressure. At the optimal parameters, you have a constant isosteric heat, and it's constant at E max and E min. Again, E max is equal to E min, and this shows you what you're looking at in an optimal case.

Now the relationship between the optimal parameters for the example we considered—I varied N max between 30 and 200 moles per kilogram. It turns out that E max and E min are independent of it. Essentially, the E max and E min do not depend at the optimal state on N max.

And when you go ahead and look at the usable hydrogen stored—gravimetric storage of hydrogen, there's a linear relation between N max and—with N max, and that's what you'd expect from the—from the UNILAN model. Now if you go ahead and think about volumetric capacity, the volumetric usable hydrogen storage is linear with respect to the product of the bulk density of the adsorbent and the limiting adsorption.

[Next slide]

Now not all materials are optimal. In fact, you may never find an optimal material. But what isotherm parameters you want to look at are good enough to meet the performance targets? Well, we know—in other words, this is a non-optimal adsorbent, but we want to just meet the technical targets for performance. We base this on the UNILAN isotherm, and again, I could use any isotherm we'd select. We just went with UNILAN because it makes a cleaner presentation for a short time.

We looked at the usable hydrogen corresponding to the difference between the charged and discharged states. The targets used in these examples are based on ultimate DOE targets for light duty vehicles. For gravimetric capacity, it's this value here, and for volumetric capacity, it's the value shown below it.

[Next slide]

Now I'm looking at the relationship between N max, E max, and E min. These parameters all vary, they vary simultaneously, you can get several groups of these parameters to meet the technical targets. Again, we start from our standard charged state and discharged state that I talked about before. Now with respect to the gravimetric target, what you're looking at here is a plot of E max versus E min, and for different values of N max, and 60, 100, and 200 moles per kilogram for N max.

And this is the relationship that E max and E min would have to have with one another for a particular N max to meet the gravimetric targets. This box shows the location of the—of MOF-5. MOF-5 does exceed the gravimetric target. Values above and to the right of these curves are—exceed the target. The values along the curve would just be at the gravimetric target. Now this is only for the material. This is not the system target that really you'd be liking to look for, but if you just look at it on a material basis, this gives you the relationship between E max and E min.

Now in terms of the volumetric target, excuse me, we—MOF-5 doesn't need it. MOF-5 has a density of about 130 kilograms per cubic meter. To meet the volumetric targets and the ultimate—DOE ultimate targets, you'd have to multiply this density by a factor of 8, or you could consider the product of the density and the N max, the limiting adsorption. That product would increase by a factor of 8. You would meet the volumetric target, and in that case, the relationship between E max and E min is shown on this—for different values of N max is shown on this graph to the right.

[Next slide]

Now stage one, and this is—we're proceeding on to stage two. Stage one, I just looked at the properties the adsorbent would have to have. Now meeting the technical targets requires more than just the isotherm properties for the—for the adsorbent itself. You actually have to consider other things. You've got to interface with the storage system. That includes heat and mass transfer, and when you heat—when you charge the adsorbent, you release the heat of adsorption, and also, there's pressure work, because you're pressurizing the system a considerable amount, and that energy has to be removed to get to the target temperatures, to your charged temperatures. Otherwise, the temperature goes up quite a bit in an adiabatic system.

Again, stage one addressed only part of the system requirements. It did not consider any kind of mass or momentum or energy transport in the system, and that all goes on when you actually put these systems into practice. The upshot is we have to consider the adsorbent and the storage system together to see if they can meet the technical targets.

[Next slide]

During charging, you have heat due to pressure work and enthalpy of adsorption, and it has to be removed, as I mentioned before. We need a sufficiently high thermal diffusivity. These adsorbent—or a sufficiently high thermal conductivity at steady state. These adsorbents, MOFs and some of the super-activated carbons, have a very, very low thermal conductivity, and that makes it difficult to remove the heat.

One way you can do this is to modify the adsorbent by adding amendments to increase the thermal conductivity, or closely space the heat transfer surfaces to reduce the thermal transport. Also, the adsorbent permeability, Don Siegel mentioned the concept of flow-through cooling. If we use that, we have to have sufficient permeability so the gas can flow through the adsorbent.

The entire mass of the adsorbent may not reach the target temperature, depending on your heat transfer configuration. We can compensate for that by increasing the total mass of adsorbent. Again, you're not going to the target temperature. You're not storing as much hydrogen. You could just add more adsorbent to store more hydrogen. The adsorbent also must be permeable enough that gas is transported to the adsorption site.

But any modifications you make, any of these modifications you make to the system that I mentioned up here, they affect the volumetric and gravimetric capacity in a negative way.

[Next slide]

What was done for the acceptability—stage two, the acceptability envelope, is that—to use a numerical model to look at the transient interaction between the adsorbent and the storage system. These are, again, transient calculations. Models include the isotherm parameters, everything from stage one; the adsorbent thermal conductivity, specific heat, density and porosity; the hydrogen flow rate, the inlet pressure—excuse me—and characteristic spacing of the heat transfer surface—as hydrogen flow rates it's important that you do flow some cooling, and the inlet pressure does tell you about the pressure work. Characteristic heat transfer spacing lets you know how far apart your safe (unintelligible) are inside the vessel, so you can reduce the thermal transport length, and it also adds to the mass of the vessel.

The differential enthalpy of adsorption is calculated from the isotherm. We have to also, when we do this stage two analysis, we have to have some system design concepts in place. Flow-through cooling, again, as Don Siegel mentioned, or cooling and heating using parallel heat transfer surfaces, like the MATI, again Don went through that a little bit. And MATI's really a super (unintelligible), if you think about it in a general term. Or cylindrical heat transfer surfaces, which are characteristic of the hex-cell configuration. They are hexes, but they can be represented reasonably well as cylindrical surfaces for the scoping concepts that we use in the acceptability envelope.

[Next slide]

Summarizing what we have so far with the acceptability envelope, and it is actually completed, we have adsorption—the viability is conducted in two stages. The first stage, we look at the amount of usable hydrogen stored by the adsorbent. That determines whether the adsorbent can possibly meet the technical targets. It also determines, that is, if we have an existing adsorbent, if we have a—we're designing an adsorbent for these to meet the technical targets. It tells you the ranges of parameters that the adsorbent must have to meet the technical targets. And we could also look at what an optimal—the best adsorbent would look like in terms of UNILAN or other isotherms.

If you meet the criteria for stage one, then second stage analysis is applied, and that determines whether the system and the adsorbent—again, you design a system to couple with your adsorbent. These systems are tailor made for particular adsorbent characteristics. It determines whether the system meeting the technical targets can be designed for an adsorbent, and it also—and if you want to go backwards, it can determine if you're going to design a system and have an adsorbent, it says what parameter ranges for my isotherms, for my adsorbent, have to be—exist, coupled with the parameter ranges for the system, have to exist to make this viable?

And these are ranges, because you can trade off between parameters. So this is why it's called an acceptability envelope. You develop an envelope of acceptable parameters for a storage system to meet the technical targets.

[Next slide]

And that concludes my talk. Thank you for listening, and if there's questions, certainly address them to myself or to the people that—in the—on the webinar, and we can make sure we address them.

Ned Stetson:
Thank you, Bruce. This is Ned again. Again, I'd like to thank all of you for listening to the webinar. We will take just a couple of questions here, but you can submit your questions, and again, we will be posting the presentation and a recording of the webinar on the website here in the next week or so. So again, we'll take a couple of quick questions here. I've got two here for you, Troy. The first one is just a clarification. On slides 8 and 24 in your material property tables, you give units for activation in kilocals per mole. Is that the correct unit, or is it really kilojoules per mole?

Troy Semelsberger:
It's kcals.

Ned Stetson:
It is kcals. Okay. And then a second question for you, Troy. You shared all of the DOE targets, so what is the current status of the storage technologies, and how are they measuring up to the targets?

Troy Semelsberger:
So for our systems, with ammonia borane and alane, alane falls shy of meeting the gravimetric target, along with ammonia borane, but with respect to all of the performance targets, we're really good with gravimetric, volumetric. The big problem that we're facing is the regeneration efficiency and fuel costs. So what I can do is I'll tack on hopefully some of the spider charts at the end of the slide to be published, and I'll actually talk with you, Ned, whether that's feasible or not. So they have a general idea of where we're at with respect to ammonia borane.

Ned Stetson:
Okay. Thanks, Troy. Now a couple of adsorbent questions. I guess first, Don, the adsorption of the MOF-5, is it induced by heating or induced by pressurization?

Don Siegel:
Both. So we're using both a temperature swing and a pressure swing. I think the idea is to try and get as much hydrogen out initially. So if you start with a full system, you want to I think use a pressure swing to go as far as you can down to say five bar, and then to get additional hydrogen out, we can—we can heat. I think we want to avoid heating if we can, because then the cooling load when we go back to the filling station to refill will be higher. But in principle, the capacities that I've mentioned involve both pressure and temperature swings.

Ned Stetson:
Thanks, Don. And adsorption question, I guess it's for you, Bruce, how would the analysis change for stage one if the adsorption temperature was increased significantly—excuse me—without compromising N max?  

Bruce Hardy:
Okay. To increase the final state temperature? Is that what you're saying?

Ned Stetson:
Change to stage one if the adsorption temperature was increased significantly without compromising N max.

Bruce Hardy:
Okay. Well, okay, N max—N max won't change. N max is the absolute maximum that you can store under any condition. But what would happen is if you increased the adsorption temperature when you—for the charged system, you'll be storing less hydrogen, and it really depends on the shape of the curve. The E max and E min give you—basically, these curves run up and break. They come up—rise quickly and then turn over.

What you want to do is make sure that you—when you go from your charged to your discharged state, you have a sufficient—you cut across the break of the curve where the curve turns rapidly downward towards zero as you go toward very high temperatures and low pressure.

The upshot is that—the bottom line is increasing the temperature would reduce the amount of usable hydrogen.

Ned Stetson:
Okay. Thank you, Bruce. And I think we're going to take one more question, and this is actually for all four of the speakers, meaning Troy, Kriston Brooks, Don Siegel, and Bruce. So today we don't have any storage materials which allow you to meet all the targets. What are the key material parameters which need to be improved for a system to meet all the targets; i.e., what key material properties need to be improved the most over current materials? So Troy or Don, do you want to take the first shot?

Troy Semelsberger:
Okay. I can go. So looking at our system designs and things like that, I mean, from an onboard standpoint, it would be the material capacity, and that material capacity that allows you to have pre- and post-dehydrogenation fluidity. Those are the critical parameters I would say that require first addressing before attacking, say, impurities. So material capacity in my opinion, with no phase change, making it facile to move on- and off-board, are absolutely critical.

Now off-board, I think it's regeneration efficiency. But I don't think you can attack regeneration efficiency until you have a material that's viable. So it's kind of a catch-22. But if I were a materials researcher, I would attack the material capacity issue with facile properties both pre- and post-dehydrogenation.

Don Siegel:
So for the adsorbent system, well, you know, this is a question that Bruce is trying to answer, so I think he should weigh in as well, but personally, I would start with the operating temperature and the volumetric capacity, and those are not independent properties. So we can make the MOF or the material itself have a higher density, but at the system level, one of the things we found is that, you know, the insulation associated with keeping the system cold is a non-negligible component of the system volume.

So those have some compounding effect, where if I can raise, you know, the operating temperature, I'm going to get back some performance at the system level on volumetric that may not directly come from changing the density of the adsorbent material itself.

Bruce Hardy:
Right. I agree with Don completely. And what I said about raising the adsorbent temperature making the—making it store less hydrogen, that's for existing adsorbents. The holy grail of this thing is to have an adsorbent that stores near room temperature, again, because you have a long dormancy, and you don't have to have this insulation as Don pointed out.

But also, when you have a—if you expected a pressure swing, what you want to have is if you look at your adsorbent curve, you'd like it to be like a step function, where you go to the maximum amount of hydrogen adsorbed, it's almost zero hydrogen adsorbed over an increment of pressure that you want to accept.  Say if you want to, say, oh, 60 bar down to 5 bar, you want somewhere between 60 and 5 bars, for the curve, the amount of stored hydrogen, to just basically drop off to zero, so you can recover all that hydrogen that you stored through a pressure swing, but at atmosphere—but near ambient temperature.

Ned Stetson:
All right. Thank you, Bruce. So unless any of the speakers want to weigh in any more, I guess with that we'll end the webinar. So Alli, I guess it's back to you.

Alli Aman:
Yes. I just want to just take a minute and thank everyone for joining, thank today's speakers. And just a reminder. The recording along with slides will be posted to our website in about 10 days. I'll send an email out once those post. And thanks, everyone, for joining.