A key role of materials and chemistry R&D researchers is to invert the primary function of their labs – that of creating materials from chemical structures, formulations and processes – to one of determining the inputs that will produce materials with the desired properties with minimal iteration. This process can be significantly accelerated by ‘leveling up digitally’, ultimately transforming lab performance and its impact on the business.
‘Digitally Leveling Up’ Your Lab
Digital technologies and their rapid advances are making their mark on R&D labs worldwide. Scientists now sense the potential for innovation at a new level, creating new possibilities in what their labs can deliver to the business. Lab automation enabled by digital technologies is a key element.
Chris Farrow and Mike Heiber, of Enthought Materials Science and Chemistry Solutions, posit that the transformation of R&D labs can be framed in 5 distinct levels, with measurable advances through infrastructure, data capture and use, digital processes and steadily advancing levels of automation.
Their white paper, ‘The Journey to Digital-centric Chemicals and Materials Laboratories’, explores these 5 levels, offering a view on how R&D labs can progress upwards, with ever higher performance levels. This blog takes a closer look at the 5 levels.
Level 1: The Human-centric Lab
At this level, the lab is dominated by people-centric processes. Deep experts in specific domains steer the lab in pursuit of better materials, driven by a rare intuition honed over years of research experience.
The absence of structured data capture or a systematic approach to digital innovation keeps level 1 labs highly dependent on these deep experts. These labs are not long for this world; deep experts are retiring, and in today’s research world, new experts will not reach the required experience levels fast enough to keep pace with innovation.
Level 2: The Data-informed Lab
One level up, the lab has adopted a more structured approach to data. Visualization and analysis as powered by better data capture, storage and search is already driving improved decision making in the level 2 lab. Experts are more efficient, more motivated, and more focused on performing analysis (rather than finding data).
The effects of even these seemingly small changes on lab performance are significant. And yet, in the context of leveling up the R&D lab, there is still a long way to go.
Level 3: The Data-driven Lab
A key feature of a level 3 lab is that decisions are derived directly from data, driving efficiency in experiment design and execution. The result of this is higher output, with objectives reached in fewer trials. At this level, adaptive experimental design is introduced; a mode of experimentation based on all data being well managed, with machine learning-based decision making driving more accurate predictions. The deep experts so fundamental to driving processes in a level 1 lab move instead to supervising and tuning algorithms for experimental design, optimizing time, cost, and exploration.
Here, significant advances in decision making are seen, with labs excelling at complex optimisation problems through adaptive experimental design. The fundamental shift at this level is in learning to trust the data and becoming comfortable with algorithmic decision making. One such example of this shift in action can be seen in a specialty chemical lab that cut time-to-market by months, whilst also saving $90k per formulation in costs due to the adoption of a learning-based optimization tool.
Level 4: The Transforming Lab
At this level, the journey of digital transformation of the R&D lab is underway. The vision for a digital-centric, autonomous lab is now motivating technology adoption and performance, with scientists committed to a common strategy.
The adaptive experimental design system that tentatively appears at level 3 is fully present at this level, turning data into knowledge and allowing researchers to begin seeking new data sources. A level 4 lab leverages automated laboratory equipment, physics-based simulation, and digital twins to push the boundaries of efficiency and accuracy.
Level 5: The Digital-centric Autonomous Lab
The final leg of the 5-level journey is the digital-centric autonomous lab – designed around data. Rapid decision making is prioritised, with experts now freed to focus on innovation and reinventing workflows. Disparate capabilities are integrated into autonomous systems. The fundamental shift experienced in a level 5 lab is the evolution that allows the lab to adapt quickly as business needs change.
New possibilities available to scientists and expert staff, enabled by autonomous systems, can drive revolutionary changes in how the lab engages its internal customers. This will ultimately drive a change in how the business engages external customers, including changing business models. In these labs, deep domain expertise is leveraged in search of greater overall predictive power and innovation.
Earlier this year, the 1000x Lab Initiative brought together industry leaders to discuss the vision of data-driven R&D and the possibilities such a lab could create. The vision is to increase throughput 1000x, whilst also decreasing cost per sample by redesigning experimental workflows to utilize automation, new measurements, and orders of magnitude new data. This is the level 5 laboratory.
The role of R&D management and beyond also undergoes changes at this level. Scientists and experts will be delivering never-before-seen results through the ‘leveled-up’ lab. Delivery of custom materials and chemicals in days rather than months is just one example. To deliver the business potential, the entire chain – from R&D to engineering to manufacturing to logistics – must be aligned and ready to transform.
Reaching the Highest Level of Lab Performance
Moving an R&D lab through the levels requires more than just technology. Cultural changes must keep pace, including an attitude of embracing change. Key to leveling up is frequent incremental successes, usually driven by introduction of new technologies. However, the real prize is the possibilities enabled by the now transformed lab. A focused, well planned and widely understood digital strategy will drive the transformation to a level 5 lab. Access the white paper here.
About the Authors
Chris Farrow, Ph.D., Director, Materials Science Solutions at Enthought holds a Ph.D. in physics from Michigan State University and degrees in physics and mathematics from the University of Nebraska. Chris has spent 16 years working as a physicist in materials discovery and characterization.
Mike Heiber, Ph.D., Manager, Materials Informatics, at Enthought holds a Ph.D. in polymer science from The University of Akron and a B.S. in materials science and engineering from the University of Illinois at Urbana-Champaign with expertise in polymers for optoelectronic applications. Prior to joining Enthought, he worked as a postdoctoral researcher at several institutions, where he developed improved physical models for organic electronic devices that connect the complex materials microstructure to semiconductor device physics.
Scattered and siloed data is one of the top challenges slowing down scientific discovery and innovation today. What every R&D organization needs is a data...
Top 5 Takeaways from the American Chemical Society (ACS) 2023 Fall Meeting: R&D Data, Generative AI and More
By Mike Heiber, Ph.D., Materials Informatics Manager Enthought, Materials Science Solutions The American Chemical Society (ACS) is a premier scientific organization with members all over…
There’s a long history of scientists who built new tools to enable their discoveries. Tycho Brahe built a quadrant that allowed him to observe the…
With the increasing importance of AI and machine learning in science and engineering, it is critical that the leadership of R&D and IT groups at...
From Data to Discovery: Exploring the Potential of Generative Models in Materials Informatics Solutions
Generative models can be used in many more areas than just language generation, with one particularly promising area: molecule generation for chemical product development.
Scientists gain superpowers when they learn to program. Programming makes answering whole classes of questions easy and new classes of questions become possible to answer….
OpenAI's ChatGPT, Google's Bard, and other similar Large Language Models (LLMs) have made dramatic strides in their ability to interact with people using natural language....
You know it. We know it. NumPy is cool. Pandas is cool. We can bend them to our will, but sometimes they’re not the right tools…
For many traditional innovation-driven organizations, scientific data is generated to answer specific immediate research questions and then archived to protect IP, with little attention paid...