TDP: What Went Wrong
I am still working on an essay on butanol, pentanol, and higher-chain alcohols, but it has been really difficult to find the information I am looking for. (But I have figured out why bio-butanol has not taken off). Regardless, I hope to complete it in a few days.
In the interim, in response to my TDP essay, I received a comprehensive e-mail from a reader who provided some very specific details on what went wrong. He included a presentation in which he referred to several slides, and I will pull those out and post them so the references are clear. I will also insert some comments in the text [like this].
I enjoy your blog quite a lot. Intelligent analysis is rare. Coupled with unbiased interpretation it is almost an unknown.
Saw your discussion of TDP/TCP. Pretty much spot on. As a chemical engineer I thought you’d be interested in some deeper insights of how the process works. This is all information that used to be available on the web, but most of it has been removed.
Start with the lecture (attached) by Terry Adams, CWT technical officer at MIT in April 2005 – best TDP technical article I know of [I have searched for an online version of this, but to no avail. Perhaps using the Wayback Machine one could locate an online archive of the original presentation]. The way I understand it, the process basically consists of two thermal treatment steps. The first step (slide #13) is a low temp/high pressure step that causes hydrolysis of all the biological material. A check of steam tables confirms that pressure is just high enough to maintain liquid water at the temperature given.
The first stage is followed by separation (slide #3).
As indicated in slide #14 they have a clever way of flashing off some of the water and then using the steam to heat the feedstock [This sort of heat recovery is standard practice in the petrochemical industry]. This is at the heart of their claims about high efficiency: the steam is condensed, so most of the water in the feedstock is discharged as liquid. Calling it distilled water, is of course salesmen talk that would make a used car salesman’s eye’s water.
But take a closer look: After separation only the “organic oil” goes to the second stage. After full hydrolysis (let’s just assume that for now) what monomers would be part of the organic oil? Fatty acids barely make it into this oil, due to the little known fact (see flow diagram on slide #11) that sulfuric acid is used to aid hydrolysis [If I had known that, I had forgotten about it. That does put quite a different spin on the whole process]. (DOE would call the first stage by another name: Dilute Acid Hydrolysis). Some fat-soluble amino acids. That’s it. (I bet you can figure out what cellulose fed to these two units would yield…) [It would interesting to see some yields on this. What I would really like to see is what they get if they threw corn in there. If their energy balances are really good - and even with all that has gone wrong they appear to be better than for corn ethanol - then I would like to see some experiments in that direction.]
Of course, CWT are master salesmen. The water-soluble amino acid and glycerol solution is not waste: it is a wonderful liquid fertilizer (slide #23). Talk about taking a lemon and making lemonade…
So, the “organic oil” goes to the second stage (high temperature/low pressure) where the fatty acids are decarboxylized (to yield oil) and some of the amino acids are deaminated and decarboxylized to yield who-knows-what (slide #15, point 2).
You raise the question of how on earth did CWT get their cost estimates so wrong. Well, a large factor in that would be overestimating yield (and per extension efficiency). CWT has long claimed that TDP has an energy efficiency of 85% (heading slide #12). Right there you smell a skunk. Now the dirty details.
The mass balance, slide #11 [posted earlier], shows that CWT probably did not take the CO2 that results from decarboxylation into account. This causes them to overestimate fuel production. You can easily do the calc’s I’m sure, but it is spelled out here. Apologies for the format, got mangled when they changed their format [That thread was a very good discussion on this issue; perhaps I will pull it out, reformat it, and post it at some point].
The energy balance, slide #12, does not include the energy present in the “liquid fertilizer”. What, all that glycerol and amino acids contain no energy? The water vapor also presents energy lost, even if it’s not much.
The mass and energy balances actually date from a previous publication (February and March 2004), also attached. One would expect that CWT would have discovered the error in the interceding year, and corrected it. I guess they were to busy ironing out the substantial start-up problems, such as the odor issue, you mentioned.
You may have notice a subtle shift between those two breathless Discover articles. Instead of producing 500 bbl from 210 tons of waste (first article), they now need 290 tons (20 tons of it pure pig fat), or a 28% reduction in oil yield. Instead of claiming 2.4 bbl/ton of waste, it is now 1.7 bbl/ton (validating an estimate of the maximum yield of ~2.0 bbl/to). Funny thing is Appel and his team still use the 2.4 figure in their financial analysis, even when it would help their argument to use the 1.7 real number. From the second Discover article: “‘We thought we would get $24 a ton for taking the waste,’ says Appel. ‘Instead we are paying $30 a ton.’ That alone raises his production costs about $22 a barrel.” How did they get to $22? ($24/ton + $30/ton)/2.4 bbl/ton = $22.50/bbl. Using the real number would yield: ($24/ton + $30/ton)/1.7 bbl/ton = $32/bbl. Also getting less yield would raise production cost in a number of ways, including the fact that they may be buying natural gas for heating…
So where does that leave TDP? No doubt it is not the silver bullet once claimed. None of the “anything” into oil that seduced Discover’s reporters. And costs are substantial. However, it seems like a good process for converting waste grease into liquid fuel. Much better than say biodiesel. Look at the feedstock (slide #6). How much cleaning (i.e. money and energy) would that stuff need to make it suitable as feedstock for a biodiesel plant? TDP uses sulfuric acid, whereas biodiesel uses methanol and a catalyst (usually NaOH). In terms of energy and money, I suspect TDP has the better input here. TDP yields a liquid fuel that is chemically almost identical to fossil diesel (without the sulfur and aromatics). TDP-40 can be blended with diesel in any ratio 1 to 100, without any issues. As Minnesota discovered last winter, biodiesel has some issues with cold weather. [Having worked in a Montana refinery, I can attest to the fact that winter properties for diesel are critical. I am aware that biodiesel has some problems with pour and cloud points in cold weather, limiting their usage to small blend fractions.]
The main threat to TDP, as I see it, is a process developed by Neste Oil, Finland, that I read about at GCC. Apparently this process allows an existing refinery to incorporate waste grease as a feedstock, without a radical change to the process (or a brand new SS plant). Even that process is not a slam-dunk, as I’ve seen reports of canceled projects.
So yes, you nailed it: these guys overpromised and underdelivered big time. But in terms of the big picture I give them some credit: at least we are not talking about food -> fuel (as with most of the biodiesel plants being built in Europe, proving that the food -> fuel madness is not endemic to North America). [Oh, I agree completely. It is not the process that I took issue with; in fact I do applaud their initiative. My concern was the completely willingness of so many to accept this as the solution to our energy problems. I see the same thing happening right now with cellulosic ethanol.] They probably help to advance the debate on waste -> energy quite a bit. And they do have a working plant, which is more than we can say about Washington’s next big thing, aka cellulosic ethanol. [I will probably write the same article on cellulosic ethanol in just a few years - overpromised and underdelivered. I see many parallels here.]