Why do breakthroughs in science and research that have enormous (and obvious) potential often not (or barely) go beyond the lab?

Question

Sometimes advances are made in science research that have enormous and very easily visible potential. There are two examples that I have seen recently.

One example is this newly engineered inexpensive process to convert carbon dioxide into methanol, for fuel. As the article says, "Making methanol from carbon dioxide, the primary contributor to global warming, would both reduce greenhouse gas emissions and provide a substitute for the fossil fuels that create them." This has huge potential, and it seems this should be of great excitement to researchers and activists alike working in this field.

Another example is this development of four strategies that represent partial but significant progress into the fighting of general cancers. This also has lots of potential in its field, and clearly is more advanced than many methods used today.

These two advances are by no means the only ones of such significance that I have seen, and there are surely much more that I have not seen, yet none of us see these go beyond this stage.

Why do breakthroughs in science and research that have enormous (and obvious) potential often not (or barely) go beyond the lab, and what makes them different from those that do become applied in the real world?

Answer 1

They are just flashy press releases. They're really not as cheap (including both capex and opex), easy, high yield, resistant to catalyst impurities, or scale-up able as written.

I've been seeing these flashy press releases on thin film fuel cells and the like (methanol production, energy production, F-T synthesis, biodiesel, switchgrass etc.) since at least the 90s. The current press releases look amazingly similar to the same press releases from back then, with little evident improvement in the field. Or realization of (or shame for) the past press releases that went nowhere. Just a new assistant professor and another press hungry R1 school putting out more sizzle, sans steak.

And don't get me started on "nano". That was already max hype in the 90s and has basically gone nowhere commercially (i.e. outside the realm of ivory tower careers). And it still seems like a hot area for NSF funding and the like. And it's a field that really attracts the ego scientists and self-promoters.

Answer 2

Because technological developments and scientific discoveries are processes at very different size and time scales of a development often under very different boundary conditions (lab vs. real world surrounding) of an in series producable technological product.

On average it often consumes even for easy transformable high-technology concepts and ideas 5-10 years to produce a prototype that can be used to optimize a ready-to-use technology in industry.

To measure the status-quo of such a development the "technology readiness level" was conceived and is widely used in academia (research reports, funding calls,...) and industry to qualify and estimate necessary time, personnel and financial requirements to achieve a distinct readiness level or decide if a concept is in the stage to transfer it from academic R&D to industrial realization and product development.

Economical requirements are not the key factor to realize a technological concept up to the final TRL level, mainly the societal use and profit (for example, nuclear fission energy/technology has long term much higher cost for society due to nuclear waste and it's correct disposal than the revenue of the companies selling energy with it).

From an economical perspective it is known that most of the high-tech startups fail. If this is due to a unknown/misestimated TRL level of the development and too early technological realization or just a bad idea or bad economical project managment would be interesting to know. But as this failure rate of high-tech startups is pretty high, likely this is not well known or statistically studied by anyone.

When you see that the number of patents is growing exponentially but money spent on R&D rather stagnates or grows slower, a lot of breakthroughs and concepts cannot be testified for readiness at all practically.

Answer 3

I am a graduate student in chemical engineering and will try my best at answering this question in the context of my field.

The short answer is that an industrial process that uses the most advanced technologies is not economical.

There can be several reasons for this.

1. The present value of the products generated from the process over the lifetime of the plant are not greater than the investment required to get such plant up and running

2. The process would not function at an industrial scale

In this example, of a catalyst that reduces CO2 to methanol I can give a few possible reasons why this process might not work on a large scale.

• The catalyst is poisoned easily and is expensive/impossible to regenerate
• The concentrations of CO2 required for the catalyst to function are higher than those found in industrial conditions
• The product poisons/destroys the catalyst
• Water or other common impurities in effluents from CO2 producing power plants would destroy the catalyst
• Extremely high temperatures, or low temperatures or high pressures are required for the catalyst to function
• The catalyst is really expensive
• Other products are made during the process that render the methanol unusable
• Oil prices are really low, making alternative fuels less attractive investments.
• The reaction rate is too slow.

Answer 4

One example is this newly engineered inexpensive process to convert carbon dioxide into methanol, for fuel.

The reason in this very specific case (and probably many other cases) is economics.

Carbon dioxide is a gas. In the clean energy economy of tomorrow, carbon dioxide is available only if (a) stored or (b) when wind/solar power is not plentiful and thus fuels are burned for energy. The process requires an energy input, so you can use it only when wind/solar power is plentiful. There is a mismatch: you create CO₂ when clean energy is not plentiful and need it when clean energy is plentiful, so you need gas storage.

I'll assume here that the process cannot operate on 400ppm of carbon dioxide in the Earth's atmosphere currently but requires concentrated carbon dioxide. There are processes such as photosynthesis that operate on 400ppm of carbon dioxide in the Earth's atmosphere. In fact, there is millions of years old technology called "tree" that converts carbon dioxide into solid fuels using sunlight by doing photosynthesis. I know this because I'm a forest owner and own tens of thousands of these things called "trees".

If you're going to store significant amounts of CO₂, you already have gas storage capacity and thus are not restricted to liquid fuels. You can store gaseous fuels as well. In particular, you can store methane (natural gas). As lot of natural gas has been used, there is a huge number of depleted natural gas fields that have the demonstrated ability to store methane for millions of years. So, for this methanol process to compete in the marketplace, it must displace water electrolysis (invented in 1800) and Sabatier reaction (invented in 1897). Both of them are well-known technologies and when used together, convert carbon dioxide into synthetic clean methane.

If I'm an investor with focus on clean energy and considering whether to fund a methanol production process, I'm primarily looking at these things:

• Longevity. The equipment must withstand use for tens of thousands of hours at the very least.
• Energy efficiency. The equipment must show energy efficiency gains over water electrolysis + Sabatier reaction to be able to successfully compete with synthetic methane.
• Usefulness of output. Current cars do not run on methanol which is a highly corrosive fuel. Electric cars and hydrogen cars are emerging. We may not ever have a large fleet of methanol-powered cars. In contrast, cogeneration and combined cycle gas turbines have a very high energy efficiency and can already burn methane. I'd much rather invest in methane production than methanol production for this very reason. There is a huge amount of methane powered electricity generation capacity already installed, and we probably have more methane powered cars than methanol powered cars.
• Cost of output. It must successfully compete with biofuels, for example, for internal combustion engine cars that do not excel in energy efficiency and are soon replaced by plug-in hybrids at the very least or fully electric cars.
• Investment cost of equipment. The equipment must be cheaper than electrolysis cells and Sabatier reactors.
• R&D cost. Commercial companies for example have invested in electrolysis cells and already paid the R&D. Does it make sense to pay further R&D for methanol production?
• Other alternative investments. For example, if you can store gas, you may be able to store hydrogen (*) which requires no CO₂ capture, if fuel cells + electrolysis cells will ever become cost-efficient. Hydrogen has better energy density by mass than methane, but poorer energy density by volume than methane. Yet, it may make sense to have a hydrogen economy for the reason that no CO₂ capture and storage is required with hydrogen.

I'm sorry to say that as an investor I'll put my money into wind power, solar power, inverters, maximum power point tracking equipment, hydropower, electrolysis and forest. (In fact, I have already put my money into electrolysis by investing into an electrolysis cell manufacturing company, into hydropower, into forest, into MPPT equipment, into inverters, into solar power and into wind power.)

(*): there are challenges in storing hydrogen because it's a very light atom so it can diffuse easily, and hydrogen can embrittle steel

Answer 5

I would like to provide an answer from the biological/medical perspective.

The process of developing a drug for a disease is extremely long, complicated and expensive. This means the whole process typically takes more than a decade and over 1 billion USD. This is a major investment for a pharma company, and some initial "scientific breakthrough" in the lab is just the first of many steps needed to make a new drug. For example, basic scientific research cannot be done on humans, so it is often done in other biological systems such as mice. However, mice are not humans, and what often happens is that a treatment that works great in mice either does not work for humans or has some very serious side effects which make it unusable. So even if some initial work seems promising, it can mean that the chance of it becoming a drug increases from 2% to 10%.

Another thing to consider is that even for breakthroughs that do become drugs, it takes so long for a breakthrough in basic research to become a drug, that it may seem like nothing happened. Only a decade or two later the impact will be evident.

Finally, what may be presented as a "scientific breakthrough" is sometimes not really a breakthrough. This can happen for several reasons, for example the university PR which wants to attract attention from the general public, or the scientists overselling the results in order to publish it in a high-impact journal.

Ultimately, science advances in small steps, with work of many scientists building on the work of others. Each of these can be considered small breakthroughs, and they push the boundaries of knowledge a bit further. Big scientific breakthroughs are rare.

Answer 6

These are all good answers, but one thing that hasn't been mentioned is what can unkindly be called Fraud.

Or as they say in research circles "The inability of third parties to replicate the results". To say this has become pervasive in scientific research is an understatement.

So in addition to all the above reasons why a "breakthrough" cannot be commercialized, today we have to acknowledge the very high probability that it in fact doesn't exist at all.

Answer 7

Eliezer Yudkowsky has written a book, Inadequate Equilibria, on this subject.

The table of contents describes chapter 2 as How, in principal, can society end up neglecting obvious, low-hanging fruit? and if chapter 2 is the how, chapter 3 attempts the why.

I'm not an economist, so there's no way I can explain this as well as the book does, but it's a principle of stack exchange that I try not to leave link-only answers. The gist of these chapters can be summarized by the quote Usually when things suck, it's because they suck in a way that's a Nash equilibrium.

According to this theory, there would need to be at least two broken things about the system preventing their immediate, widespread adoption by single actors. In other words, two sets of independent actors would need to coordinate through a nest of conflicting incentives and inadequate information.

In the case of cancer care, you can observe pretty easily that effective, shared information is hard to come by. Patient privacy laws are a net good, but they do make some kinds of research harder. Since healthcare in general is very resistant to change, convincing evidence is both hard to come by and very much required for anything to happen.

Such lucky coordination doesn't happen very often or very fast.