Science!

One of the strange things about campaigning for public office is that your life is necessarily reduced to a set of soundbites. They aren’t wrong — but they are far from complete. And the sheer volume of public engagements and limited time for one-on-one interaction makes it difficult to ever answer detailed questions with the richness they deserve. This is especially true for answers requiring detail and nuance which may put half of any given audience to sleep.

With that in mind, here’s an effort to answer one of the questions I’m often asked but rarely able to answer fully. “What did you do when you worked as a scientist? And what about that experience is relevant for a sitting member of Congress?”

7 Years a Scientist

During my senior year in college while contemplating career options, I knew two things. One, I loved my major (Molecular Biology and Biochemistry), and especially my classes in genetics, molecular biology and organic chemistry. Two, unlike everyone else in my major, I didn’t want to be a doctor. So I started cold-calling CEOs in any company I could find that seemed like they might employ people like me. (The hubris of youth!) The handful who took my call all agreed that an undergraduate degree was necessary, but insufficient. I’d have to go to get a graduate degree if I wanted to ever get a meaningful job in the field. And I‘d also have to get some practical experience in a working lab before I’d be able to provide any value to their firms.

That advice led me to my first job at Tufts School of Medicine in the lab of Dr. Barry Goldin. He and a colleague had discovered a new species of bacteria that was closely related to the bacteria used to make yogurt. They discovered that when added to rat food as a probiotic, it appeared to alter the rats’ production of certain hormones which had separately been found to reduce cancer risk. Dr. Goldin had focused specifically on the ability of that dietary supplement to affect markers for colon and breast cancer.

For the former, we were still in the animal trial phase, and my job was to manage a fairly large population of lab rats on two distinct diets, then count and measure the comparative frequency and size of colon tumors in each. For the breast cancer research, we had moved onto human trials and had a large number of women who had volunteered to adjust their diet for our research and then come into the lab to provide us with urine, blood and fecal samples.

[A brief break from science talk to ask you to consider my plight as a single man in Boston in my 20s, trying to answer the “so, what do you do for work?” question as I was working the weekend Boylston St. bar scene. “In addition to euthanizing rats, I collect fecal, urine and blood samples from pre- and post-menopausal women!” was not as effective as you might think.]

While I learned a lot about how to “do” science during those two years, I also learned something bigger about the process of science itself. The science you read in textbooks tells history in reverse. “Here is Darwin’s theory of evolution, and this is how he figured it out.” “Here is Marie Curie’s theory of radioactivity and here is how she figured it out.” But science in practice moves in the other direction. The day-to-day work of scientific research is a much messier story of false starts, failed hypotheses, poorly timed equipment malfunctions, unsolved puzzles and dead-ends. It is simultaneously mundane and fascinating, tedious and precise. And it’s a job that depends on lots of young, smart, very methodical people to put in the hours of work to collect the gigabytes of data that the Ph.D. managing the lab can hopefully look at one day and discover some previously unnoticed pattern. I enjoyed my colleagues and the work but after 2 years, I was ready to move on.

So I proceeded on to graduate school at Dartmouth College to get an M.S. with Dr. Lee Lynd. Lee has, for years, been at the forefront of research into cellulosic ethanol. Ethanol is of course the stuff that makes beer interesting. It is also a fuel oxygenate, octane booster and increasingly important part of the U.S. automotive fuel network. In America, it is made primarily from corn, and has raised a large number of environmental and economic concerns due to the fertilizer-intensity of corn as a crop and the economic challenges innate to mixing our food and fuel feedstock chains.

The process of making ethanol is not complicated. Humans figured out how to do it over 2,000 years ago and the basic idea is still the same, whether used at the fanciest French winery or the dingiest basement hooch still. Get sugary water. Add yeast. Cover. Wait for the bubbling to stop. Voila. The problem from a fuel perspective is that sugary water (whether in the form of grape juice, molasses, corn mash or malted barley) is pretty expensive per unit of energy. That, in a nutshell is why it’s hard for traditionally-produced ethanol to be cost competitive with gasoline.

Lee’s insight was that sugar in nature exists in multiple forms. It can be in free molecules (think table sugar). It can be strung together in long “polysaccharide” chains as starch (think potatoes). Or it can be organized in a more tightly bonded crystalline structure as cellulose (think wood). Plants naturally make those compounds as a way to store sugar. Fungi and bacteria that live on the forest floor (or in compost piles) have evolved to break those complex compounds down into sugar that they then digest for energy. And in general, the more complicated the chemical structure, the cheaper the raw material. So if we could develop fermentation systems that used the bacteria in compost pits instead of food grade yeast, we could dramatically expand the potential for ethanol use, substantially without the economic and environmental challenges from corn-based ethanol.

My job? Try to get the reactors to work. Specifically, while we had good systems in place to turn wood into ethanol, the concentrations we made were too low to be economically recoverable. A predecessor in the lab (Sunitha Baskaran) had shown that our limitations related to the fact that our bacteria made one molecule of acetic acid for every molecule of ethanol. Since the acid eventually lowered the pH of our reactors, we had to add caustic so as not to kill our bacteria, either in the form of potassium hydroxide or sodium hydroxide. Sunitha’s insight was that while we could boost production by adding caustic, eventually the sodium and/or potassium levels themselves became toxic. That became the starting point for my own research.

Specifically, I set out to try and replace sodium and potassium hydroxide caustic solutions that are normally insoluble. That would allow us to inject large volumes of (solid phase) material into our reactor that would dissolve only as necessary to neutralize the acid, but would then precipitate back out as solid-phase carbonate compounds. The chemistry here is less important than the mechanical impact: what had been a reactor where liquids entered, were mixed and then poured out was now a reactor where liquids and solids entered, sloshed around together and then had to be removed without concentrating all the solids at the bottom of the vessel. A biology problem became a chemistry problem became a mechanical engineering problem.

Did it work? Sort of. But the mixed-phase fluids problems proved much harder to solve in the real world than on paper. Feed lines clogged and pumps jammed a lot more often than they used to, and every repair introduced potential contamination into fermenter. And every contaminated fermenter cost 2 weeks of data. (Lee’s lab has since solved this problem by genetically modifying the bacteria to knock out their acid-producing pathways.) However, I gradually became more interested in another problem — namely, that while the low cost of cellulosic materials make them attractive as a feedstock for fuel markets, the conversion processes are notoriously inefficient, for the simple reason that 40–60% of a wood chip isn’t cellulose. It’s like trying to make cream out of whole milk — you can do it, but unless you’re also doing something with the skim milk, it’s very wasteful.

That led me to spend the bulk of my master’s thesis developing detailed and sophisticated computer models of the entire ethanol production process, including the ancillary infrastructure required to generate the heat and power to run those plants. I was able to show that in a truly integrated energy plant we could use the “wasted” part of the wood to generate the heat and power needed to run the overall operation. Moreover, if we assumed that we used next-generation heat and power conversion technologies (much as we were for the ethanol production technologies), an ethanol plant could be a net producer of electricity to the power grid in addition to producing fuels, creating the possibility for significant improvements in overall fuel efficiency.

After graduation, I was hired by Arthur D. Little, where I joined a group that had just developed a technology to convert gasoline into hydrogen that would fit inside a car. The team that had developed that technology was being spun out into a separate business, leaving ADL with a sudden need for people with expertise in emerging power and fuel conversion technologies. My first assignment was to build a computer model of combustion chemistry that included a solid phase catalytic process. I had to take a crash course in C++ programming but the combination of multi-phase systems and computer modeling of chemical processes picked up perfectly from where my master’s thesis had left off. Subsequent projects including evaluations of manufacturing costs for next-generation fuel cell technologies, refinement of rotary (“Wankel”) engine technologies, comparative assessments of various early stage metal-hydride, lead and lithium-ion battery technologies and a series of assignments for the Dutch government to evaluate various carbon-free fuel chains they could potentially invest in to ensure the lowest cost, cleanest future for The Netherlands.

Gradually, my role became more strategic and less scientific, and when I transitioned from ADL in 2000 to become an entrepreneur I effectively put my ‘hard science’ career behind me. But not without picking up several lessons that have informed my approach to every problem since.

1. Always question your hypotheses. We all suffer from confirmation bias, but the most efficient way to ask the most interesting scientific problems is to work under the assumption that your conclusions are wrong. Try to prove they’re wrong and you’ll uncover all sorts of interesting lessons, not least about the (usually narrow) set of circumstances where your theory is correct. But if you only try to prove your theories correct, you will often be blinded to the truth.
2. Any problem worth solving involves a lot of grunt work. There’s no glory in filling in data tables, or numbering tissue samples, or reconciling accounting records. But if those jobs aren’t done well, the big picture isn’t available. Either do the grunt work yourself, or else surround yourself with good people who will. And make sure they know that you know you depend on (and appreciate!) them.
3. Opportunities sit inside of problems. Running from problems is emotionally easy. Diving into them is hard — but more often than not, that dive turns up fascinating opportunities.
4. Finally, knowing things is fun. Digging into a problem, peeling the onion, and identifying the root cause is the only way that we’ve ever solved problems of any complexity. Far too many businesspeople, politicians, and failed scientists get caught up in soundbites and unquestioned paradigms. We succeed as a society when we dig in.

Sean Casten is the Democratic nominee for Congress in Illinois’ 6th district. To learn more about Sean, visit www.castenforcongress.com.