Role: Design lead
Currently, the road to materials discovery is a long and expensive one. On average it takes 10 years and 10 million dollars to discover a new material. As the world faces increasing challenges such as climate change, plastic pollution and food shortages just to name a few, the need for novel materials to address these challenges is more pressing. IBM Research is working on a new initiative that will accelerate materials discovery with the goal of shortening the time to one year at a cost of one million dollars.
This goal can be achieved by reimagining a materials discovery platform, which consists of several tools that will enable this paradigm shift. Our team’s goal was to work closely with researchers to develop an easy to use interface and validate market value.
Designed at IBM Research
Currently, the road to materials discovery is a long and expensive one. On average it takes 10 years and 10 million dollars to discover a new material. As the world faces increasing challenges such as climate change, plastic pollution and food shortages just to name a few, the need for novel materials to address these challenges is more pressing. IBM Research is working on a new initiative that will accelerate materials discovery with the goal of shortening the time to one year at a cost of one million dollars.
This goal can be achieved by reimagining a materials discovery platform, which consists of several tools that will enable this paradigm shift. Our team’s goal was to work closely with researchers to develop an easy to use interface and validate market value.
Role: Design lead
Designed at IBM Research
What does materials discovery look like?
First we needed to understand what materials discovery even was, the current process and who does it. We conducted dozens of interviews with researchers, both internal and external, to paint a picture of the current as-is flow. Getting this right was essential in understanding the jobs to be done at each step, tools used and identifying pain points along the way.
Identifying the best problem to solve
As our interviews progressed we created concept cards to help facilitate discussion. Each card addressed a recurring paint point from past interviews and proposed a solution. At the end of each interview we asked researchers to rank each card based on if the proposed solution would be helpful to them. Concepts were either ranked or discarded. This exercise was valuable in validating potential paths forward as well as letting go of ideas that had little value. When our research phase concluded, we decided that simulation was the most promising path forward and the best fit for our team.
The democratization of simulation
The research team whom we partnered with had a clear focus on the value of simulation and how to make it more useful to more people. They had two distinct personas they were addressing in their approach, the computational and the experimental chemist. Computational chemists use models to simulate properties and reactions either to anticipate what may happen in the lab or explain the results. Experimental chemists on the other hand are in the lab conducting the real experiments. While the two disciplines complement each other, there are several issues. Since each speaks their own language there is a communication barrier between the two. Computational chemists are rarer, expensive and often with a lengthy queue of work that needs to be done. Simulation is often out of reach for most experimentalists. The current tools are hard to use and often inaccessible. Our problem to solve was how can we create a tool that allows experimental chemists to easily run routine simulations while freeing up their computational colleagues to work on more complex and interesting problems. Our solution was to develop an intelligent simulation tool where a experimental chemist could run a simulation with only a few clicks, keeping as much under the hood as possible.
As easy as buying a plane ticket
Through several rounds of iteration and user feedback, we built a MVP that could run a simulation with only a couple of clicks. The UI was kept as minimal as possible focusing on the job to be done but also offering explanation and guidance at key points in the process. One potentially tricky interaction was the tool required an experimental chemists to choose a simulation method. To make this easy, the tool suggests what the best methods are determined by the chemist’s budget. Depending on the task they need to achieve, they can select the fasted method (less accurate) or the most accurate method (slowest) and so on.
We’re excited to continue to develop this critical piece of a future material discovery platform, and hopefully have a real impact on issues facing the world today.