The Need for Fuels Produced Using Synthetic Biology

Among the most promising short term applications of Synthetic Biology is biological production of liquid fuels.  But beyond the technical and economic attraction of the project, the reasons we require progress in this area are manifest; diversification of fuel sources thereby reducing dependency on imports, improving air quality, reducing greenhouse gas and particulate emissions that contribute to climate change, eliminating the present coupling between biofuels and food crops, and carbon sequestration.

Bio-era is in the middle of scheduled briefings in Asia, the U.S., and Europe describing the present state of biofuels markets and associated technologies, and these trips, along with recent headlines concerning commodities prices and future fuel demands, have helped clarify the story in my mind.  Below I outline some of the factors in play:

Carbon and other Greenhouse Gas Emissions: The amount of water coming off Antarctica and Greenland scares the crap out of me.  It's true that this isn't my professional specialty, but I have been following the literature on polar ice mass and movement for a decade.  The news is just getting worse.

The present coupling between biofuels and food crops creates upward pressure on food price inflation and reduces (or eliminates) the economic incentive to produce biofuels: Ethanol demand has pushed up the price of corn, and in the U.S. politically motivated trade barriers to Brazilian ethanol derived from sugar cane threaten to keep corn prices high.  Palm oil is presently trading at historic highs, and at a ~30-40% premium to finished diesel, but this is actually driven by food demand, primarily from India and China.  I am a simple physicist by training, rather than a sophisticated economist, but given the increase in food demand I don't see the price coming down even with increased supply.  This puts anybody planning to refine palm oil into biodiesel completely underwater for the foreseeable future. 

China (and India) will require increasing resources over the coming decades: More on this in posts to come.  The numbers are mind boggling.

Ethanol is by no means an advanced biofuel; from both a technical and an economic perspective ethanol is a backwards biofuel: The future is all about producing biofuels that are high energy content (not ethanol), are not water soluble (not ethanol), can be easily integrated into the existing gasoline and diesel distribution infrastructure (not ethanol), and require minimal, if any, initial changes in engine technology (not ethanol).  The average age of an automobile in the U.S. is now at least 10 years (depending on who is counting, and how), which means engine technology turns over very slowly here.  It is faster in other countries (2-3 years in Japan, if memory serves), but this dramatically influences the speed with which new fuels can enter the market.

You don't want to be long on petroleum in ten years:

First, despite a greater than 10% annual growth in auto sales in China, petroleum demand has evidently plateaued due to increased biofuel blending.  I'm not sure I completely believe this yet, but it is an interesting assertion.

Second, three companies are already out in front with funding to use both traditional metabolic engineering and synthetic biology to produce microbes that churn out biofuels:

LS9 is "Developing Renewable Petroleum biofuels: new, clean, and sustainable fuels that fulfill our long and short term energy needs. Derived from diverse agricultural feedstocks, these high energy liquid fuels are renewable and compatible with current distribution and consumer infrastructure."

Synthetic Genomics, Craig Venter's shop, just announced a partnership with BP aimed at using organisms and genes found in subsurface hydrocarbon deposits to develop "cleaner energy production and improved recovery rates".

Amyris Biotechnologies recently received $20 million to develop direct microbial production of liquid biofuels.  Amyris, in particular, is well positioned to make some serious headway.  The company website suggests they are well on their way to making both butanol and biodiesel  (or more likely a precursor to diesel?) in microbes.  In an article in Technology Review, the new CEO, John Melo, says the company has already developed a metabolic pathway to produce a fuel equivalent to Jet-A.  This is particularly interesting given the recent announcement by the U.S. Air Force that it will replace at least 50% of its petroleum use with synthetic fuels by 2010.  In an article by Don Phillips, The New York Times is reporting that, "The United States Air Force has decided to push development of a new type of fuel to power its bombers and fighters, mixing conventional jet fuel with fuels from nonpetroleum sources that could eventually limit military dependence on imported oil."  At the moment, the immediate plan appears to utilize a synthetic fuel produced using natural gas, but anybody who can crack the aviation biofuel nut has immediate access to a 3.2 billion gallon per year market in the Air Force alone.

So how long is this all going to take?  Amyris CEO Melo mentions they hope, "To make a Jet-A equivalent with better properties on energy and freezing point with a $40 barrel cost equivalent by 2010 or 2011".  That's faster than I was expecting, but I find the time scale highly credible.  Below is a figure with data drawn from Jay Keasling's recent presentation to the UC Berkeley faculty senate on BP's investment in the Energy  Biosciences Institute.

Isoprenoid_yieldThe data represents a roughly billion-fold improvement in yield over 6 years.  (I've called this "pre-synthetic biology improvements" because the data is the result of applying fairly traditional metabolic engineering techniques, rather than the combination of Biobricks.  This is by no means a critique of Jay Keasling or his teams at UC Berkeley or Amyris, but rather a simple contrast of methodology.)

You would be hard pressed to find examples of that magnitude of improvement in any human industrial process over any 6 year period, but that is exactly what is possible when you turn to biology.  Moreover, the complexity of the isoprenoid pathway is probably about the same as you would expect for producing biobutanol or a Jet-A equivalent.  This is why John Melo is bullish about making progress on biofuels.  Given that Amyris is evidently already on the path towards butanol, diesel, and aviation fuel, five years is by no means an overly optimistic estimate of reaching commercial viability.  Note that this doesn't mean Amyris takes over the liquid fuels market overnight.  It can take decades for new technologies to make progress against existing infrastructure and investment.

But assuming Amyris, or any other company, is successful in these projects, it is worth considering first the resulting impact on the liquid fuels market, then more generally the effects on structure of the economy as a whole.

The economic considerations of scaling up direct microbial producing of biofuels are fundamentally and radically different than those of traditional petroleum production and refining.  The costs associated with finding a new oil field and bringing it into full production are considerable, but are also so variable, depending on location, quality, and local government stability, that they are a poor metric.  But a very clean measure of increasing gasoline and diesel supplies is the fractional cost of adding refining capacity, presently somewhere between US$ 1 and 10 billion dollars for a new petro-cracking plant, plus the five or so years it takes for construction and tuning the facility for maximum throughput.

In contrast, the incremental cost of doubling direct microbial production of a biofuel is more akin to setting up a brewery, or at worst case a pharmaceutical grade cell culture facility, which puts the cost between about US$ 10,000 and 100,000,000.  Pinning down the cost of a biofuel production facility is presently an exercise in educated speculation, but it is more likely to be on the low end of the scale suggested above, particularly for a fuel like butanol, which, unlike, ethanol, is not soluble in water and therefore does not require distillation; it can simply be pumped or skimmed off the top of the tank in a continuous process.  Beer brewing presently occurs at scales from garage operations bottling a few liters at a time to commercial operations running fermenters processing thousands to many millions of liters per year.  Thus, once in possession of the relevant strain of microbe, increasing production of a biofuel may well be feasible at the local level, thereby matched to fluctuations in demand.  Microbial biofuels could therefore be an excellent initial demonstration of distributed biological production (PDF warning).

In the end, the scalability of microbial production of biofuels depends in part on what materials are used as feedstocks, where those feedstocks come from, and how they are delivered to the site of production.  Whereas petroleum products are a primary feedstock of today‚Äôs economy, both as a raw material for fabrication and for the energy they deliver, it may eventually be possible to treat biomass or waste material as feedstocks for microbes producing more than just fuels.  But as I observed above, any biological production process  for biofuels that relies on a sugar or starch crop also used in food production will be subject to the same skewed market dynamics now playing out between food and conventional biofuels.

There are clear challenges to overcome in the years ahead, but given the progress already demonstrated I am comfortable we will find solutions with continued effort.