Discards

I generally think it’s good to be promiscuous with ideas. Move among them quickly and discuss them openly. The main reason that startups keep their stuff secret is that they’re embarrassed of stuff that doesn't work yet, not that there’s any genuinely secret sauce to protect.

In that spirit, here are a few topics we've bandied around this year without success. At this point it's unlikely we'll pursue any of them, so they may not be worth stealing. If you're intrigued or are working on something similar, please get in touch. We’ve compiled a lot on each of these and are happy to share our analyses.

‍

Polyglycine

The cheapest possible protein, producible anywhere and from a variety of feedstocks: polyglycine, made via chemical condensation of ammonia and vinegar.

Bubbling ship exhaust into the ocean

A simple decarbonization strategy for maritime transport gets a closer look. Conclusion: 5-10% emissions reduction at near zero cost, but an uphill startup journey.

Artificial rogue waves

Focusing structures would have to be large and steerable to effectively concentrate wave energy to useful power densities

All-polymer batteries

The highest possible charge rate for an electrochemical battery might require the oxidizing, reducing, and ion-conducting elements to be integrated as repeating functional groups on aligned polymer chains. Fabricating it cost effectively is still beyond modern chemistry.

Aluminothermic magnesium

There’s enough low-grade or hard-to-recycle Al out there to produce all of the world’s Mg in an aluminothermic process. A nice circular gambit, and the economics look OK. It’s not a very exciting version of the future though—we’d rather wait for someone to work out a genuinely improved electrochemical cell that makes Mg from the oxide, or perhaps a thermochemical process, “Pidgeon Mg with American Characteristics”…. 

Digesting cellulose to glucose

Cellulose is composed of glucose monomers, which can be recovered in solution with a low-cost acid treatment. But the economics can't beat $0.10/kg sugar from corn. If Coleridge’s mariner were instead lost in a forest: "Sugar sugar everywhere and not a bite to eat".

Laser pyrolysis of biomass

The faster you pyrolyze biomass, the cleaner and higher-value product you tend to make. Could low-cost lasers—the fastest way to apply heat known—help us build a new type of pyrolysis plant for low-carbon fuels? Possibly. But the implementation could be very tricky, and the improvements over existing pyrolysis processes would be marginal at best.

Nylon via bulk oxidation of unrecycleable polyethylene

Bulk oxidation of low-quality polyethylene gives short chain diacids, preparation of which is ordinarily the most emissions-intensive step in nylon production. Could we make cheaper and cleaner nylon using old plastic bags? Unfortunately, the variety of diacids you’d encounter in this process means it wouldn’t be a cost effective way to produce nylon 6-6, or other high value polymers.

Microalgae as edible biomass

Open-pond saltwater algae could provide very cheap protein and a healthy spectrum of fatty acids. It's unlikely anyone will want to eat them unless they’re highly (and expensively) processed.

Synthetic wood

The market size for exotic hardwoods doesn't seem to justify a development effort on the scale of cell-based meat, if mimicking wood to connoisseur standards is possible at all.

Long strands of DNA

As our ability to predict protein structure and function increases, synthesizing very long strands of DNA may become the key to one version of 'plenty of room at the bottom’.

Biosynthesis of unsaturated fatty acids

It doesn't look like an enzymatic pathway will be able to beat current algal processes for synthesizing ω3s and other high-value unsaturated fats.

Protecting lowland areas from inundation by injecting stuff into subsurface aquifers

Land subsidence contributes more to Venice’s flood risk than climate-related sea level rise does. And much of the subsidence under Venice is thought to result from centuries overdrawing their subsurface freshwater reserves. Could we raise seaside cities by injecting either water or mud back into these aquifers? Probably not—much subsidence is irreversible (e.g. stepping on powdered snow), and leak-off limits what you can do with ordinary hydraulic fracturing techniques.

Decarbonized cement from gypsum

Green cement with no green premium? A decarbonized variant of the Muller-Kuhne process for making CaO from phosphogypsum would use elemental sulfur to reduce CaSO₄ to CaO and SO₂. While it could be a way to make clinker without CO₂ emissions for $50-100 per ton, because the process binds the fertilizer and cement industries together, there are limits on the amount of cement it can provide: roughly 7% of world demand, offsetting only a few hundred MT of CO₂  per year. More of a a sustainable fertilizer add-on than a green cement moonshot.

Sinking biomass in anoxic seas

The depths of the Black Sea are famously anoxic. Fresh water from Eurasia pours over the top of colder saline water that enters through the Bosphorous straight. The result is extreme density stratification, and remarkably O₂-poor waters that leave classical shipwrecks in near-pristine conditions for millennia. While we think that rafting and sinking biomass in the Black Sea would be a cheaper carbon drawdown than burying sterilized waste in plastic bags, ultimately we don’t love either solution.

Ship sulfur

We thought that weather-adaptive tactical marine diesel resulfurization might be an effective way to offset 2-10% of anthropogenic warming at zero or even negative cost. It seemed like an extension of the logic behind out contrails project; contrails was about undoing ongoing harmful geoengineering, this would go step one step further by reinstating a previous “business as usual” helpful geoengineering entitlement. However, after long contemplation we concluded that while the International Maritime Organization’s desulfurization policy might have been somewhat misguided, sulfur simply isn’t a good low-altitude geoengineering material, and putting it into fuels deliberately would probably have measurable costs in human health, even if you tried to do it only in remote open-ocean areas.

Floating islands

Rafts of biomass were interesting for at least three reasons: 1) “New land” to offset sea level rise. Land is good, people can live on floating biomass islands (Uros island, the Phumdis of India), and if they’re left wild terrestrial habitats are generally more biodiverse than ocean ones. 2) Carbon sequestration. Woody biomass decomposes much more slowly in water than on land. Rafts of logs in water can be exceptionally long lived, especially in lower oxygen waters (c.f. the Spirit Lake log raft or the Great Raft that clogged the Red and Atchafalaya rivers for ~1000 years).  3) Albedo. Biomass rafts can hold snow and weather to a white color. Three weak hooks, no actual market need, and no compelling technical advantage means this one’s not for us.

Fossil nitrogen

Rocks adjacent to coal beds can contain elevated nitrogen levels, which are a byproduct of fossil decomposition. Unfortunately, the high cost of extracting rock borne nitrates means that we are unlikely to make economical fertilizer this way. 

Earthsuits

In a warming world, will there be situations where it make sense to cool individuals, rather than whole spaces? Generally not, we think.

Thermochemical sulfur oxidation

We reevaluated an old process for thermochemical hydrogen production from high temperature heat, the sulfur-iodine cycle. We speculated that if we removed the most challenging step in the process, sulfuric acid thermolysis, the cycle could become a one-way process for co-producing sulfuric acid and hydrogen. However, our analysis revealed that even the ‘easy’ steps of this chemical looping scheme were never going to be cost-effective.

Decarbonized steel from sulfide ores

Refining of hard rock ores produces SO₂: 4FeS₂ + 11O₂ > 2Fe₂O₃ + 8 SO₂. By electrolytically oxidizing the SO₂ to H₂SO₄, we could cheaply produce more than enough H₂ at the mine site to reduce the resulting Fe₂O₃ to pig iron—a tantalizing route to low-carbon steel for certain deposits. In the case of chalcopyrite (Cu, Fe & S) ores, the incentive would be to run the sulfur oxidation process galvanically, in order to produce H₂SO₄ and electricity rather than hydrogen. Unfortunately, upon closer look, no real-life sulfide roasting process makes SO₂ at high enough purity for a sulfur-depolarized electrolyzer to run effectively.

Natural nuclear reactors – turn them back on?

Two billion years ago, natural nuclear reactors lit up deep underground under Gabon in West Africa. Groundwater dripped down into natural uranium ore deposits, which at the time contained about 3% ²³⁵U. As the water infiltrated the ore and began reflecting neutrons, the ore bodies ‘went critical’. Based on analysis of the decay products, one particular “natural nuclear reactor” produced hundreds of kilowatts for hundreds of thousands of years with a duty cycle of about 3 hours; water dripped in, the nuclear reactor turned on, the water boiled off, and the reactor turned back off. There aren’t any more natural nuclear reactors today, since the concentration of ²³⁵U has naturally decayed to about 0.7% in all earthly ores (and indeed, probably in all ores across our stellar neighborhood). However, are there things we could do—graphite/heavy water slurry injection, down-hole enrichment—to make one again? The result could be “enhanced geothermal power” worthy of the name.

The most likely start up path would have you rounding the "edgy" parts off this idea until you were marketing an underground PWR-- something other people are already doing.

Microbial products from electrosynthetic feedstock

At least 1000 startups aim to replace common ag- or animal- derived products with alternatives fermented from i.e. glucose. It might be cheaper and better for the earth if the starting material were produced agriculture-free: formate, acetate, or a simple alcohol. While our ‘food without agriculture’ thesis area is as strong as ever, and while we think something like this will emerge eventually, we haven’t seen an opportunity to inject anything game changing yet. 

Long duration grid scale energy storage in commodity chemical streams.

Almost any commercial oxidation process (how we get nitrate, chlorine, sulfuric acid, hydrogen peroxide, ethylene oxide…) could be electrochemically coupled to a bifunctional hydrogen evolution/oxygen reduction electrode to store almost arbitrary amounts of energy in the form of a commodity chemical stream. It’s a breathtaking prospect that just might help increase the penetration of renewables on the grid, but it might not be applicable at quite the scale we’d hope for.

Post-fossil phosphate production

In a net-zero world, we’ll lose the lowest-cost sulfuric acid we use to produce phosphate fertilizer. But other sulfur sources will pick up the slack long before other acids are used or a direct approaches to sulfur recycling become economical.

Superconducting power transmission

We looked into bottlenecks for rolling out both conventional (MgB₂) and high temperature (YBCO) superconducting grid infrastructure, and decided that without an epochal improvement in critical temperature or cryocooling technology, superconducting transmission lines won’t  (and shouldn’t!) girdle the earth any time soon.

Bringing uncertainty into the DICE model

The DICE model for climate risk assessment has received a lot of attention, but also a lot of flak. Perhaps most controversially, the model does not take into account uncertainty about climate impacts. We looked into reworking DICE to use optimal portfolio theory to find the optimal carbon tax, but didn’t see how the result would be any more informative than adjusting DICE’s damage function.

Protein templates for condensation

We can predict protein domains that will act as effective seeds for disordered ice crystal growth. Is there a bridge from that to a hyper-effective bio-based cloud seeding material? In one vision, crops would release cloud-seeding pollen or pollen-like material—farms that make their own rain.

Maps of the atmosphere: signals of opportunity

Microwave links to satellites might provide surprisingly accurate atmospheric tomography for water vapor, which could improve climate models, weather forecasts, and enable better contrail avoidance. It would be complex to operationalize, however, and it’s uncertain who your customer would really be.

Maps of the atmosphere (platforms of opportunity)

Better maps of atmospheric CO₂, CH₄, CO, H₂O and aerosols would have undeniable utility in plotting the future of our climate. Is there any way to get paid to measure them more accurately than we do now?

Space solar

If you can concatenate a few optimistic assumptions, you can show that geosynchronous PV arrays could supply cheap and location-flexible baseload power to earth. The best version of this idea uses NIR lasers to illuminate conventional solar farms at night or in the evening. But you'd probably just be larping a startup with ground-based demos for years. The real start-up cost could be staggering.

Dendrite-free Li batteries

Clever management of concentration gradients could dramatically reduce dendrite formation on Li-metal anodes. While a 1D model suggests that this strategy alone would be enough to stabilize a Li-metal anode for many cycles, in reality it’s one of many strategies that will have to be tried (and are being tried) in concert for our metal anode dreams to come true. 

Decarbonized shipping

We considered a few proposals for electrified transoceanic shipping. We concluded that incentivizing shippers to go slower might be a better investment than most proposals for new technology.

Planetary thermostat

In a future world with plentiful low-cost energy, would it make sense to develop an active cooling system to radiate excess planetary energy to space? Certainly not, it turns out. A highpoint of this analysis is its reduction to absurdity of solar geoengineering via land albedo changes.

Atmospheric war observatory

Nation states rely on seismic stations and various radiological measures to determine whether other nation states are messing around with nuclear weapons. We could easily build a similar monitoring system for conventional-munition conflict using NOx-sensitive climate-observing satellites to look at anomalous oxidizer use. However, there are other more convenient means to detect ongoing open warfare.  

Electro-fission of soot

There's a moment during combustion when the polymerized aromatic compounds that will become soot coalesce into liquid droplets. Could these droplets be fractured by electric fields in turbofan engines in order to change cloud-seeding properties? Yes, but you'd need to redesign your whole engine for it.

Dry ice depth charges

There is a fair (5-15% emissions reduction) decarbonization opportunity in using counterflow heat exchange to cryogenically separate CO₂ from the exhaust of an LNG tanker. Then simply drop the solid CO2 overboard where it sinks and stays for thousands of years. It’s a similar- size opportunity to bubbling ship exhaust directly into the ocean. However, the cost would be much higher, some of the objections (e.g. replacing one form of pollution with another) would be similar, and the modification would only be applicable to the ~5% of the fleet powered by LNG.

Nitrogen fixation with metal nitrides

It lets you reduce N₂ without Haber-Bosch, and the metal nitride can then donate nitrogen directly to a carboxylic acid to yield a very soluble and bioavailable primary amide or even amine compound. But the overall efficiency is even lower than plasma- based nitrogen fixation-- there was no getting around the large free energy loss in the initial nitride formation.

Boudouard reaction for partial combustion

Like methane pyrolysis, but for heat rather than hydrogen. What if we partially combust fossil feedstock to H₂O and CO, and then disproportionate the CO to solid C and CO₂? A low O₂ version of this process may also yield H₂. Unfortunately, like pyrolysis or carbon suboxide formation, this approach is lower emissions, but ultimately less attractive for all the free energy it leaves on the table. You’d have to produce a high value solid carbon product and there’s no a priori guarantee that you would.

Anion redox anodes for alkali metal batteries

Of course we all want reactive metal/metal oxide batteries. Al-air would be twice the energy density of diesel, Mg- and Li-air nearly as good and easier to engineer. In the course of studying the progressive oxidation of Li for Li-O₂ batteries, Ju Li’s group at MIT started impregnating Li in ordinary cobalt oxide intercalation cathodes. It’s a good platform for studying the progressive reaction Li-Li₂O-Li₂O₂ you’d expect in Li-O₂ batteries. But if you look carefully at their data, you see that it could also make a formidable (~5-8MJ/kg) and extremely cycle-stable positive electrode on its own.

Fine wines

Regional fine wines might be the first culturally and economically significant artifacts we lose to climate change. Should we publicize their plight, or work to systematically predict which regions are most vulnerable so that we can invest now in bottles likely to increase in value? Probably not worth it. The disappearance of prized wines is self-publicizing. Ultimately, their preservation benefits few, but their extinction could galvanize many.

Making more land

The debate over geoengineering always seems to miss a few things. For one, the extensive geoengineering we already do— hosting artificial organisms (livestock or crop monocultures) on half the habitable land of planet earth, corralling the majority of terrestrial freshwater flows, seeding half the clouds in the northern hemisphere. But the debate also misses a lot of the bigger geoengineering we probably would do, if we were really optimizing our planet for human or animal welfare. Most prominently to me would be making new land. The land is much richer than the sea from a biodiversity standpoint, and also much more useful from a human standpoint. Hear Ishmael in Moby Dick: “Yea, foolish mortals, Noah's flood is not yet subsided; two thirds of the fair world it yet covers” If we’re thinking evenhandedly about geoengineering options, maybe we should brainstorm a way to remove some of this water, deepen the ocean so the water covers less of the earth, pump seawater into subsurface aquifers to raise coastal cities, or at least dredge up some new earth to replace a sunken Florida Key or two. For better or worse (perhaps better 🙃), all the suggestions here sounded like different versions of Atlantropa or Project Plowshare and we dropped it.

Industrial transamination

Many visions for the future of the food system include protein produced by microbes. In these visions, the nitrogen for protein production would be supplied as we supply it to plants, in the form of ammonia, nitrates, and urea. However, microbial conversion of these nitrogenous compounds to amino acids could be challenged by toxicity and competing organisms. One alternative to microbial protein synthesis would be to  engineer microbes to overexpress pathways for transamination, which would let you supply a ready-made amino acid into machinery for amino acid interconversion and peptide assembly. Producing glycine as a feedstock, for example, would let you take the chemistry farther in the high-rate, high conversion chemical context before making the handoff to biology.

Sulfur-burning vehicles for SAI

A fleet of planes for stratospheric aerosol injection would likely leave as much CO₂ and H₂O as SO₂ in the stratosphere. At the same time, sulfur is a viable fuel with energy density comparable to coal. Should we develop high altitude engines that can burn S to SO₂ instead?

‍