Solving real world problems was the starting point for Dr Linden Schrecker’s journey in science, taking him from a chemistry degree to interdisciplinary doctorate and now founder and CEO of a spinout from Imperial College London, called SOLVE.
“As a child, I originally thought I wanted to do biology or botany because it felt like the way to discover things that you could bring into the world that would be useful. But, to me it seemed chemistry provided that path between the creativity that you might find in biological sciences and the rigour of mathematics or physics,” Schrecker explained.
“One thing I concluded as an undergraduate is that while reading academic papers, you realise that if you follow the reported process there’s maybe a 60% chance that you can repeat the results. Things are not as repeatable as we might think. So, this took me down a route of wanting to do more data driven automation-based research, which shaped my PhD,” he said.
With an industry backed research programme as part of the EPSRC CDT Next Generation Synthesis & Reaction Technology underway, at Imperial College London, Schrecker was on the road to realising that link between creativity and solving real world problems.
“The initial idea behind my research project was the self-optimisation of chemical reactions - automatically screening conditions for a reaction until we find the optimal set of conditions for that reaction. But I soon realised that there was a mismatch between what industry was saying it wanted and what they needed.”
This mismatch led to a conversation with Schrecker’s industrial sponsor, BASF, and supervisors confirming Schrecker’s thinking. “What industry is looking for is the optimal conditions for their reactions at scale. They really don’t care too much about optimising at gram scale as self-optimisation would get you. So, to optimise, what we need is enough data to predict what the outcomes will be at scale,” said Schrecker. “But to optimise one has to decide what optimal means, market conditions such as energy costs and regulations change, therefore what’s optimal changes. What we really need is a lot of data underpinning a flexible model.”
With numerous parameters required to optimise a chemical process, Schrecker set about thinking how best to collect these data sets, which led him into the realm of transient flow. Transient flow allows accurate data collection more than 20 times more efficiently than batch experimentation. “With transient flow it’s possible to collect data in a way that is efficient, repeatable and covering multiple parameters including temperature and concentrations, and monitor conversion, yield, kinetics,” explains Schrecker. Building on this methodology, which had been documented in academic papers but was commercially unrealised, Schrecker found that he was able to screen solvents in an efficient manner and get useful information based on different solvent parameters.
“We realised that we could patent this work and that there was also a commercial route for this. BASF saw the value that it could provide for their business as well.” The SOLVE process utilises machine learning to choose only the experiments that will provide the most useful information.
“We’ve been able to make some real improvements on the machine learning side since our Chief Scientific Officer, Jose, joined us following the spinout,” Schrecker notes. “On some parameters, we’ve already halved the time it takes to collect information. Using transient flow coupled with AI and the volume of data that can be collected is what really sets the technology apart, de-risking process development decisions and improving efficiency” says Schrecker.
Starting with solvents made a lot of sense, he said. “The solvent space is shrinking. European Union and US regulations are getting stricter year on year, the need for alternatives is huge. Following some industry surveys across sectors such as agrochemicals, fine chemicals and pharmaceuticals we realised that finding solvent replacements was hard, and even harder was collecting data. There really isn’t very much process data collected or stored to solve these issues, especially the quantity and quality we can achieve with our transient flow techniques,” says Schrecker.
This ability to help customers better understand their processes gave birth to the company’s name. “Applying what we had found on a commercial basis on process parameters beyond solvents had also not been done before and we realised that we could offer even more value across a wide range of parameters. The project name, SOLVE Solvents, changed to reflect the greater ambition in our company name: SOLVE Chemistry,” says Schrecker.
Taking SOLVE from a research project to a business happened quite quickly. “We formed the company during April this year, with the process of spinning out from Imperial College taking around a month. Fast compared to the usual timescale of six to 12 months,” said Schrecker. Early funding has come from a well-respected UK venture capital fund called Creator Fund as well as Chemovator a 100% subsidiary of BASF. Schrecker points out that while Chemovator is a subsidiary of BASF, the multinational chemical company has no shareholding in SOLVE. “We have a great relationship with BASF, we’ve been able to use our technology on a solvent replacement test case with BASF India, but we are also free to sell into other markets.”
“We really want to be a facilitator for the chemical sector. With the data we can collect we can better inform businesses on how they can best build their production facilities providing a route for a faster time to market. We could provide better information to help on the regulatory front and make an impact on the sustainability of the chemical industry,” says Schrecker
Further reading on science and innovation
- Life sciences and clean energy core to new Industrial Strategy
- Six new centres aim to grow bioeconomy research
- Biopharma company gets financing to advance work against biological threats
- Decarbonising concrete: Low Carbon Materials secures £3m growth funding
- Battery breakthrough could help EVs go further
Visit Blog