Episode 80: Decentralizing Chemistry

This episode explores the innovative approaches to decentralizing chemistry, aiming to enhance the efficiency and scalability of chemical processes outside traditional large-scale facilities.

1. Decentralized chemistry seeks to replicate large-scale chemical production efficiencies on a smaller, more localized scale.
2. Innovations like flow reactors and electron spin control are central to advancing chemical process efficiency and precision.
3. DARPA's projects include creating portable systems for on-site pharmaceutical production and advanced recycling technologies.
4. The integration of AI and machine learning in chemical synthesis is poised to revolutionize pharmaceutical manufacturing.
5. Cross-disciplinary approaches are crucial, combining chemistry with process engineering, systems theory, and informatics.

1: Introduction to Decentralized Chemistry

Dr. Vishnu Sundaresan discusses his journey to DARPA and the foundational concepts behind decentralized chemistry. Tom Shortridge: "Decentralized chemistry, essentially, is about achieving large-scale chemical production efficiencies at smaller, localized scales."

2: Innovations in Chemical Processing

Exploration of new chemical processing technologies that leverage electron spin to control reactions, aiming to reduce unwanted byproducts. Vishnu Sundaresan: "Using electron spin as a control variable could revolutionize our approach to chemical manufacturing."

3: Practical Applications and Future Visions

Discusses the practical applications of decentralized chemistry in military and civilian contexts, including on-site drug manufacturing and advanced material recovery. Vishnu Sundaresan: "Our goal is to enable the production of necessary chemicals and pharmaceuticals close to where they are needed, which is crucial for military operations and remote areas."

1. Embrace interdisciplinary approaches in chemical research to integrate new scientific insights and technologies.
2. Consider the environmental and economic benefits of localized chemical production to drive innovation.
3. Stay informed about advances in robotics and automation for potential applications in chemical synthesis.
4. Explore the role of electron spins in chemical reactions to enhance processing efficiency and product purity.
5. Engage with emerging regulatory frameworks for decentralized pharmaceutical manufacturing.

In this episode, Dr. Vishnu Sundaresan (https://www.darpa.mil/staff/dr-vishnu-sundaresan) from our Defense Sciences Office highlights several technology programs designed to precisely control chemical processes to enable distributed, small-batch manufacturing of chemical products while retaining efficiencies of large-scale industrial production. Colloquially calling this portfolio “decentralized chemistry for everything,” the concept aims to shift the paradigm from a few centralized production facilities producing medicines in large batches and requiring a costly purification process, to direct manufacturing of pure pharmaceuticals via desktop printer-sized machines that would create — at the push of a button — doses of a variety of medicines whenever and wherever needed. Such a revolutionary capability — if successful — would circumvent brittle international chemical supply chains and would serve military members deployed in remote locations as well as benefit rural civilian communities.

People

Tom Shortridge, Vishnu Sundaresan

Companies

DARPA

Books

None

Guest Name(s):

Dr. Vishnu Sundaresan

Content Warnings:

None

Unknown
Coming to DARPA is like grabbing the nose cone of a rocket and holding on for dear life. DARPA is a place where if you. Don'T invent the Internet, you only get a b. A DARPA program manager quite literally invents tomorrow. Coming to work every day and being humbled by that.

DARPA is not one person or one place. It's a collection of people that are excited about moving technology forward. Hello, and welcome to voices from DARPA. I'm your host, Tom Shortridge. On this episode, we're speaking with Doctor Vishnu Sundaraysan, program manager since 2020 in DARPA's defense sciences office.

Tom Shortridge
But your introduction to the agency came well before that. I came to the United States as a graduate student in 2003, January, and I knew of DARPA, and I always aspired to look at DARPA's innovations in the past while I was a student back in India. But then when I came to the US, I had the opportunity to work on a DARPA funded program, which ultimately. Became my PhD thesis. That was my first exposure to DARPA as a performer on the DARPA program in the first year of my graduate school.

Vishnu Sundaresan
The topic was on developing biologically inspired materials, essentially to take plant proteins and to see if we can build engineering devices, such as actuators and energy conversion devices. We had some hopes that it could. Work, but there were a lot of. Fundamental challenges with materials that did not. Make it work as good as an engineering material should.

As all Doppler programs go, that was not the only outcome. What we realized while working on that program was you could take biological materials. And interface them with synthetic materials and. Exactly use the same form of interactions that biology uses, which is ion transport, and build sensors that can work at scale. That became my career focus after I finished my PhD.

So that failure led to a huge. Portfolio on biologically inspired sensors and biosensors and neuromorphic computational devices. That led to four or five PhDs in my group. So then, fast forward to 2020. You're an associate professor at the Ohio State University in mechanical and aerospace engineering, and you had the opportunity to come to DARPA as a program manager.

My interest to come to DARPA was. To focus on what I would call as structural computing or structural computers. DARPA has had a history of investments. In that space, and the most fundamental. Challenge that we always had was how could we bring sensing, actuation, and computational modalities all within a structure?

I had enough academic work in that space, but nothing that could have a DARPA like impact. So I wanted to come to DARPA. To actually build a portfolio in that space. There were other priorities for the agency at the time and still continues to date. Therefore, I had multiple portfolio options, and I picked one that I thought could.

Have the most impact in the agency, which ended up being decentralized chemistry for everything. Decentralized chemistry for everything. What exactly does that mean? Simply what it means is you want to be able to do chemistry or chemical processes with the efficiency that can. Be brought to bear, that can only exist in very large scale.

So when you look at most chemicals that are manufactured, that is an integral. Part to our lives, starting from the synthetic garments we use, to petroleum, to pharmaceuticals, to mining, to recycling, they're all done in a centralized factory. And economics of scale works out in such a way that it is profitable. Only if you do this such large scales. But more often than not, there is not enough market justification to build a huge plant to serve a niche need, which is pretty much Dod's need.

So DoD's push, at least in my portfolio, has been to see how we can achieve the variety, the variability of products that we want at the scale we want, the right scale we want, but the same cost efficiencies that can be had only at very large scale. Essentially, I want to have my cake. And eat it too. So under this decentralized chemistry for everything portfolio, I've managed four programs. The first one was a program that I inherited called make it, which was focused on developing robotic tools and automation for synthesis of fine chemicals, and then specifically later applied to manufacturing pharmaceutical drug products.

Unknown
The second program, in decentralized chemistry for everything, was focused on inorganic chemistry, primarily for recycling critical elements from electronic waste. So this program is called recycling at the point of disposal. So we can do recycling in a distributed manner in a decentralized manner at multiple locations, more preferably at the points of collection of e waste. The third program is spin controlled chemical process engineering. And here we are going after the fundamentals of chemistry to see if we can use electron spin as a reaction variable to control a chemical process.

And then the last program is equipe pharma. Here again we are developing metrology tools and therefore informing the creation of a new regulatory framework, if successful, for distributed manufacturing and agile manufacturing of pharmaceutical drug products. So again, under these four programs, we are covering organic chemistry, inorganic chemistry, fundamentals of almost all chemical processor chemical reactions, and lastly, but not the least metrology to monitor these chemical processes and reactions. For those of us who didn't go past high school chemistry, why is chemistry so important? Chemistry pretty much makes or breaks our world.

Vishnu Sundaresan
For example, let me ask you this question. How did you come to work today? I drove a car. In your car, you've got hundreds or. If not thousands of plastic components that are present that make it work.

The fuel that you use, the synthetic oil that you use in your car. For example, it's fully made by a chemical process. Now, when you want to drive a chemical process, you just don't want to make something, but you want to make it as efficiently as possible with as low a cost as possible and with as minimal a waste as possible. So to do that, just chemistry alone isn't enough. You need to bring in concepts from process engineering, systems theory, control theory, to drive chemical processes.

So essentially, what you now see is your high school chemistry now becomes a lot more complicated and sophisticated with inputs from so many other disciplines. And then you can add in instrumentation, data, and informatics theory on controlling a. Chemical reaction at scale. All those things now makes it a very multidisciplinary problem. We've been trying to do that in many different ways, and I believe at this point we've almost hit the saturation of what is actually possible in controlling.

A chemical reaction, at least until the launch of a program that I launched about a year and a half ago. Scope. It stands for spin controlled chemical process engineering. When you think of a chemical process. These days, we typically use pressure, temperature.

Unknown
Flow, rate of reactants, and concentration of reactants to control the outcome. But for a chemical bond to be formed, the electron spins of those reactants will have to be opposite to each other when they form that bond. Up until now, we never knew how to control that. But in this program, we are exploring through some new fundamental insights that we have to see if this could even be a plausible route to control a chemical reaction. When you think of a chemical reaction.

Reactants combine together and form products. But in most reactions, the formation of. Products is accompanied by byproducts. More often than not, the byproducts are either undesirable or toxic, or reduce the efficacy of the main product that we want. But the process of separating byproducts from products is very expensive and also very slow.

It is therefore important for us to see how we can completely eliminate the formation of byproducts so that we don't have to. That extra separation step that becomes a. Important cost driver for manufacturing things at a much smaller scale. And scope program fits in this picture because we are now looking at completely shutting down the formation of a byproduct using electron spin as a control variable. How does spin prevent byproducts for a.

Vishnu Sundaresan
Chemical bond to form? If you go back to your high school chemistry, we see that electrons live in their orbital shells, and then an electron could have a spin up or. A spin down for a chemical bond to be formed. The electrons that come together to form that bond, one should have a spin up and the other one should have a spin down. That is how all reactions happen with current paradigms.

We don't really control the spin of that reactant as it is reacting together. But with the new approach, we're now. Trying to influence the spin states of all those radicals that are present in a chemical reaction. So that would now have spin up and spin down for only those that. Want to come together, and everything else would always have a spin up, both of them, or a spin down state, so that bond will never, ever be formed.

So we are going back to high. School chemistry, but again, applying that, using. Fundamentals of material science to change chemistry forever. And if this program is not complete. Yet, we don't know if it's going to be successful yet.

But if it is successful, it will. Become a chapter in high school chemistry. And that is the level of impact this program could have, and obviously just so much more for the DoD and the warfighter community. So if we can completely create a new discipline called quantum electrochemistry, this program will be one of the programs to do that. How does decentralizing chemistry fit in with the defense mission?

For example, when we send our warfighters. To fight off in far, far away. Places, producing pharmaceuticals is an important problem statement. And we don't need the dosage level that is needed for the general public. We just need just the right dosage.

For the platoon or the squadron. The number of troops that are deployed. In that faraway place, their scale, small scale becomes important. Navy is transitioning some of our programs. Previous programs on chemistry, to put a pharmaceutical manufacturing box on the deck of a ship.

So for such things, you need to. Be able to make things at a small scale, but still have a reasonable. Efficiency, if not efficient, if a big pharmaceutical company. So that is why producing things at small scale is extremely important. And again, you can also think of.

Other very sophisticated lubricants that DoD might need. You don't need to produce them in gallons and barrels, vats like you know others would need. But for the right application, you may. Just need the right quantity for achieving. The end use application.

Producing in small scale is important. So thinking of the heilmeier catechism, specifically the why now component? Have there been recent breakthroughs or tools that you've seen that make you think now is the time that this is possible? There are a couple of fundamental innovations. That have happened in this space.

One is called a flow reactor, in. Which instead of reacting things in big. Batches, you let these reactants, chemical reactants, flow through a tube, and then you. Control the temperature, the pressure concentration of. Reactants in that small tube, and you.

Therefore control the outcome of the reaction. So here you're producing things in a. Very small control volume, and therefore efficiencies. Can be brought to bear. But there is one paradigm, and then again, you can throw in automation, machine.

Learning, advanced instrumentation, metrology, to again understand what is reacting, what is being formed, and therefore control the chemical reaction with some efficiency. And the next thing which I said is electron spin. If you then bring an electron spin, and as another reaction variable that can. Now be controlled, you're now looking at. Efficiencies that are just simply not possible.

Using any other paradigm. You talk about decentralized chemistry as a portfolio. Another of your programs is called recycling at the point of disposal, or RPod. How does that fit in? So, typically, when you think of recycling.

E waste, electronic waste, recycling, it is always considered to be a very inefficient process. And the reason for that is we. Simply borrow techniques from mining industry to do recycling of e waste. When you look at mining industry, they go after the biggest market that they can service with their technology. And the two extraction techniques that almost.

Every recycling company that we have around us uses. They're called hydrometallurgy, or pyrometallurgy. One that uses water, lots of it. Or heat, again, lots of it. And when you look at e waste, electronic waste, or any critical material, rare earth elements, these are present in very.

Very small quantities, hundreds of parts per million. These elements are present in very small quantities. And again, there are so many of them that we cannot use existing technologies. To recover those elements. The program recycling at the point of disposal, what we wanted to achieve was.

To see if there's a way in. Which we can play with thermodynamics and reaction kinetics to just put in one unit of energy and then fractionally recover. All these elements that are present in very small quantities. Just like how crude oil refining works. You heat up crude, you volatilize as.

Many things as you want, and then. You fractionally condense one after the other. So in the RPoT program, the most. Successful outcomes that we've had are to use sulfidation the reaction between the chemical. And sulfur, or the reaction between, say.

The material and chlorine or carbon, volatilize them and then fractionally condensing them so that we can now just put in. One unit of energy and recover ten different products. So now we are completely changing the economics of recovering these materials, instead of having to just recover the biggest component that we can from a feedstock with. Recycling at the point of disposal. We have now developed the technique to.

Recover even milligrams of materials that are present and, say, 100 grams or a kilogram feedstock. That was never considered possible before, but. Now we've shown that it is plausible. And then performers even went further ahead to even show that it is actually possible at scale, to recover 500 grams. Of something from tens of kilograms of.

Feedstock, which is, again, a miracle, if. You think of it in the industrial process perspective. We're now trying to see where else it could be applicable. For example, we are exploring the feasibility to recover gallium and germanium from, again. E waste to see if that could meet the needs of DoD's use case.

So, again, getting back to the decentralized chemistry portfolio, here we are using a. Novel chemical route to do chemistry at. A small scale, so that you can. Imagine these r pod boxes to be. Just like a red box kiosk, if you will, that can do recycling at a local store.

Tom Shortridge
Now, once you have your materials, whether they're coming from an RPOD system or whether they're what we more traditionally think of as chemicals, are there novel ways to do things with those? I inherited a program called make it. Which was to showcase the feasibility to. Use AI ML tools to control chemical synthesis, essentially manufacturing of fine chemicals, pharmaceutical products. We were very good in showing that we can make a single hardware and in a single hardware, produce at least three or four pharmaceutical drug products.

Vishnu Sundaresan
You put in all the raw materials, you have a hardware. That hardware can flow the reactants through. Various pathways, and depending upon what output you need, will produce that said output. And if you want to change it, you can again push a button, it'll. Produce a different drug product.

It's like a desktop printer, which is. Bigger, but it can make different drug products. When we wanted to scale this technology. Up to the market, it turned out. That because of regulatory restrictions, the market was not ready for the technology.

For example, if a Walgreens or a CV's pharmacy has a drug printer like this, they could just have some key starting materials. They can actually make multiple drug products. But if they make it, they'll have to meet FDA's regulatory approval pathways to sell it in the market. So the market in this case was not ready for the technology, and your. Equipa pharma program is picking up from that point.

With Equipe Pharma, we're establishing qualification processes for agile pharmaceutical manufacturing. Performers will set up pilot manufacturing sites that can showcase this for multiple drug products and simultaneously develop a wealth of data on how this process can be. Quality monitored, quality controlled, and more importantly, regulated. So to do that, they're going to. Collect data and showcase that using local.

Data, they can approve the drug product. In addition to that, they're going to. Also develop a broadly agreeable process markup language. Think of it as your HTML Hypertext markup language that we use for browsing the Internet. They'll develop a process markup language that.

Tells a computer where the reactants are going, how they are going, and combined with the data, they'll send it off to a future server. And this software that resides on this. Future server could accept this process markup. Language data, and the data itself coming. From the manufacturing hardware and electronically approve the drug product.

I would make an analogy here to. Say it's probably how we do our taxes now. In the good old days, we're doing. Paper taxes just like we're doing our taxes today. I wanted to be an electronic file that is completely transmitted with electronic fingerprint.

And just like how our tax are approved electronically, drug manufacturing could also be approved in an agile pharmaceutical manufacturing paradigm. Obviously, we're a long way out, but if Equipa Pharma is successful, what do you envision that could look like 1520 years from now? Pharmacies can have a single box that. Can make drug products and squish local regional shortfalls. And that is a very serious problem facing not just the DoD, but also.

The nation at large. As we know, in the past summer we had a significant shortfall of chemotherapy. Drugs, and there's an ongoing shortage of. Drugs for the last three decades or so. While we always think that we're developing technology for the DOD, more often than not we see that every technology the DARPA touches goes to the benefit of mankind at large.

So that ultimate goal is always there in every, at least in my mind. I'm sure it's in every program manager's mind, as big of an impact as you possibly can. I mean, certainly see there's a potentially huge impact here. We've just done a couple episodes on Elsi, ethical, legal and societal implications of technology. How are you looking at those considerations in your portfolio.

So when you're now thinking of making. Pharmaceuticals or drug products in a decentralized. Manner, you're opening a Pandora's box of things, regulatory things, legal things that can come into play here. So it is extremely critical for us to engage with our LC partners, identify early on what are all those challenges. So that we can absolutely be sure.

On how the technology is being developed. And how responsive and how mature it. Is when it meets the market, so that we have all those safeguards in place. We have had some initial conversations, but again, it is important that LC is a critical component for everything that we would want to do with pharmaceuticals or chemistry. Aside from the LC community, are there any other groups at the agency that you've been working with that people might not be as familiar with?

One shout out that I would like to give in. DARPA is the commercial strategy office. That office is at its infancy and has everything in DARPA. DARPA is an experiment, CSO is an experiment within DARPA, and commercial strategy office. Is a very unique entity that can critically support almost every DARPA program.

And I've been very fortunate to get their help. So with make it. When we had the technology and we were shopping that technology around, retail investors cannot invest in this technology. It takes tens of millions of dollars to get this technology to the marketplace. And almost every time when we went.

And spoke with a VC that, you know, has any appetite to even hear us out, the first question they would always ask us is, yes, your technology may be ready for the market, but. The market is just not ready for the technology. Such a feedback would not have come to us unless we engaged with the. Market VC community, in this case through commercial strategy, to therefore take that as an input and design a new technology solution to get the market ready to receive the technology. So such things are extremely important.

And within the LC framework, we can also look at this as a legal. Attribute for the market to be ready for the technology. And that could happen only if commercialization. And getting the technology to the market is a part of our mix. So DARPA typically develops just technology, but.

More often than not, guess what? It's somebody that has to make it for it to actually get in the hands of the warfighter. That commercialization piece is extremely critical and. Important for DARPA's mission to be successful. And any final thoughts before I let you get back to work?

I would want more people to come. Knock on DARPA doors to be a PM or to engage with DARPA as much as possible. DARPA is a very unique agency that. I'm sure everybody would have parroted this at some point. But as I'm getting ready to leave the agency in a couple of months, it strikes me more so than what I thought of it coming in.

That's a very rare opportunity to serve the nation and the public, and there's no place anywhere else in the world. To, to call it home and to. Be very proud of it. That's all for this episode of voices from DARPA. For more information on the programs we discussed, we'll have links in the show notes.

Tom Shortridge
Thanks to Doctor Vishnu Sundar Raisin for appearing on the program, to Eric Vaderbaugh for his assistance in producing this episode, and thank you for listening.