Using AI and machine learning can allow formulators to narrow down possibilities significantly, saving time for more innovation.

Source: Sebastian Duda - Adobe Stock

There is tremendous interest in the potential of Artificial Intelligence (AI) and Machine Learning (ML). There is almost an equal amount of concern over AI. There are plenty of questions surrounding AI; one question is how AI can be used for the benefit of industry. Companies are slowly seeing the potential.

What is AI and ML Anyway?

AI (Artificial Intelligence) is a broad field aiming to create machines that can mimic human intelligence, while ML (Machine Learning) is a subset of AI that focuses on enabling machines to learn from data and improve their performance without explicit programming.

Samiul Amin, professor of practice and director MS program in product design and undergraduate program director - Chemical, Environmental and Materials Engineering at the University of Miami’s College of Engineering, has been at the forefront of AI. He notes that while AI and machine learning are areas with a lot of current interest, they aren’t necessarily completely new technologies.

“AI and Machine Learning have evolved from expert systems that have been around for a long time,” Amin observed. “There have been expert systems that could guide instruments for specific measurements in R&D, in quality control, process optimization and predictive maintenance in manufacturing, in diagnosis and treatment planning in healthcare etc.

“AI and Machine Learning have been an evolution from expert systems where the performance depended on the knowledge initially programmed into it to the AI/Machine learning systems of today, where you can get much higher-level insights and guidance through continual learning from new data,” Amin noted.

“The adoption of AI has been slow, and people do not know how it can be effectively used,’ he added. “Using AI for product development, you can tune the formulation to achieve certain performance characteristics and can take it to a place that you couldn’t have achieved without conducting hundreds of experiments.”

How Coatings and Ink Companies Use AI and ML in Formulation

Specifically, AI could benefit coatings and ink manufacturers in numerous ways. Formulation is an ideal area for AI; Professor Amin has co-founded with Professor Yelena Yesha (Knight Foundation Endowed Chair of Data Science and AI at the University of Miami) a company, FastFormulator Inc., that utilizes chemical intelligence (a combination of AI/Deep Machine Learning/Formulation Science) among its disciplines.

“We recently worked on sustainable ink formulations for a large company,” Amin said. “We were researching biodegradable formulations. One area of importance is the way the ink flows. For example, with a ballpoint pen, the way the ink flows dictates the flow of ink onto the paper and can impact the color intensity and writing quality. This is determined by viscosity and rheology.

“If a company wanted to create a completely new sustainable ink or coating formulation, it would have to go back to the drawing board,” Amin continued. “You cannot have the ink’s pigments settle; they must be suspended to ensure uniform color. It is the same thing if you are spray-coating a paint; you have to ensure the coating formulation stays stable, sprays effectively and will not jam up the nozzle, and that you’ll get good uniform coverage and good color intensity.”

Amin noted that using traditional reformulation efforts are very time and resource-intensive.

“You have several components in a formulation that control surface properties, flow properties, and stability,” he pointed out. “Optimizing the formulation takes many tests, varying one ingredient at a time, until you get the right mix. For inkjet printing, you are flowing at very high speeds (high shear rate), which is really important, and you have to ensure optimized flow properties and avoid pigment clogging. AI and machine learning can speed up this formulation optimization process drastically. Instead of doing 100 experiments, you can come down to 10 or even five.”

There’s more that goes into the formulation. New regulations can go into effect, changing the availability of allowed ingredients. FastFormulator also uses AI to keep up to date there, as well as with intellectual property (IP) and patents.

“AI can also monitor any changes in regulatory or compliance,” Amin pointed out. “AI can then give you a substitute that is allowed. When a new regulation goes into effect, it’s not usually a one-to-one substitution and this becomes labor intensive. It can also help monitor intellectual space, as well as help you find similar formulations in different industries giving companies competitive advantage.”

Creating a New Paint or Ink with AI

Amin offered the example of creating a new bio-degradable ink or paint.

“With an ink or paint, you have pigments, emulsifiers, and polymers that have to be bio-degradable,” he said. “All of these components were traditionally petroleum-based, but now you want to make them all from plant or biobased sources. If I’m a paint or ink company, how do I do that?”

“There are very large number of suppliers who now have many options for bio-degradable polymers, emulsifiers, pigments, dyes” Amin continued. “So how do you know which one will give you the performance that your customers expect? This is an evolving area with new choices of eco-friendly material options coming out on a continuous basis. If you are large paints/coatings company like a Akzo Nobel or Sherwin-Williams or PPG, your customers have certain expectations from your brands. You have choices for a huge range of novel eco-friendly materials that you have to combine in the right proportions to get the expected performance. Each combination needs to be tested. That can take months, even years.”

Amin said that with AI, a formulator can make those combinations literally within hours.

“You input the ingredients into the system and it will make recommendations automatically,” added Amin. “It may give you the top five possible formulations. You still have to test the formulations, but now you have an optimized starting point.”  

Aside from chemical formulation, regulatory compliance and intellectual property monitoring, FastFormulator also can be customized to handle other tasks.

“Our main areas of expertise are AI/machine learning/deep learning together with in-depth scientific expertise in formulation/colloid science, rheology, surface activity etc,” Amin noted. “Our first product is the Chemical Intelligence model, which optimizes the right material combinations. We can also go into the manufacturing side, as you must get the processing (mixing/pumping/pipe flow) right. We have a second AI model on compliance and regulation, which gathers real-time information. The third monitors intellectual property and patents.

“We can also customize these products,” he added. “We have a subscription model, where we provide continuous updates to the models and regulations and add new ingredients. We also have AI as a service for medium to smaller companies that don’t have a lab, as we have a fully-equipped formulation lab and can custom make product formulations for them utilizing our AI/ML platform and formulations lab.”

“We have done customized ink formulations as well as working with chemical suppliers for the paints and coatings industry,” Amin said. “We do formulation development for consumer industries as well as agricultural applications. Interestingly, the base science is the same in terms of emulsions, surfactant/polymer mixtures, flow properties, surface properties and stability.”

Concerns Over AI

There are legitimate concerns over the use of AI. People are afraid of losing their jobs, and there is a perception that AI can start making decisions on its own. Amin says that is unfounded.

“Right now, there is a certain hesitancy with AI,” Amin said. “People don’t know much about AI’s capabilities. The key is to understand that AI is a tool that can help us enhance our productivity. You have to learn what that tool’s capability is. Tools can’t do anything without our guidance; I don’t see AI going out and building things itself.

“How effectively it has been used is a question – it still needs to prove its value, and it can’t show its value unless it is effectively adopted,” added Amin. “A recent study showed that only 6% of big companies have adopted AI, while others are still trying to understand what it can do. We are seeing a lot of start-ups coming into the market that also don’t understand AI’s true value. You have to use AI the right way, filtering out the noise and finding true value through bringing together a convergence of AI/ML/Deep Learning/Formulation-Colloid-Material Sciences.

“AI is a tool and needs to be adopted - there is a fear of change, and if what we are doing works, why do we need to change it? That is an issue as people are resistant to change. It’s not about fixing a process but accelerating the process. Repetitive actions aren’t the best use of our intellect. AI allows us to be more innovative. Companies need to set aside some resources for tomorrow for the effective adoption of AI. AI requires a culture change and a mindset change, and that is difficult to achieve. Education will have to play a critical role.”

Ultimately, Amin sees great opportunities for AI and machine learning for formulating inks and coatings.

“AI will significantly accelerate the time to market,” Amin concluded. “The time you’re saving is an opportunity cost and frees you up to come up with innovative products faster and more effectively.”