Harnessing AI and Automation for Advanced Particle Size Analysis 

In this interview, Paul Senior, the Product Manager within Malvern Panalytical’s micro-materials group, talks to AZoM about how to harness AI and automation for advanced particle size analysis. 

Could you provide an overview of the digital transformation journey within particle size analysis? How have digital tools and AI fundamentally changed the approach to research in this area? 

The digital transformation journey for particle size analysis has a long history, largely because analytical instruments are inherently digital, with technologies often based on data science. Algorithms process the recorded data to generate the required output. 

For particle size analysis using laser diffraction, I view it as a process with distinct stages within the digital transformation journey. 

Initially, the focus was on ‘simplification,’ particularly in data processing. Tfhe transition from mathematically involved data processing to a more user-friendly approach marked this stage. Early systems allowed users to quickly select a pre-configured analysis mode, which automatically generated the particle size distribution result. 

Next came the stage of ‘standardization.’ Digital tools and refined software offerings enabled users to create better-quality data with superior test methods. For example, AI has started to assist in the method development process, and new file types allow all relevant test parameters to be captured and shared. 

Currently, we are witnessing the emergence of a new stage, as seen with the brand-new Mastersizer 3000+. This stage focuses on using digital tools and AI to create targeted solutions for customer problems. 

What were the key factors that motivated the integration of AI and automation solutions into particle size analysis technology?  

One trend we aimed to address with AI and automation solutions was the progressive deskilling of users. Not everyone wants or has the time to become a particle size expert. Researchers face increasing demands on their time, and we wanted to help them make the best use of it. 

The growing need for extensive data has also been a motivating factor. When dealing with ‘big data,’ the old adage of ‘quality over quantity’ no longer applies. Customers need large quantities of data they can trust, ensuring both quality and quantity are met. 

Additionally, here at Malvern Panalytical, we are driven by a desire to advance the technologies used in particle size analysis. While laser diffraction is a relatively mature technique, there is still much to be achieved, and AI and automation are playing a significant role in these advancements. 

You mentioned about the deskilling of users – let’s touch on that topic in some more detail. How has the shift towards more automated and data driven methodologies helped address this challenge? 

Deskilling of users is a particular concern for companies when it comes to particle size analysis because it is vital that measurement best practices are followed at every step of the process. Sampling, sample preparation, measurement parameter selection, and data interpretation are all crucial steps. 

Without careful attention to each of these steps, the particle size result may not reflect reality. For example, you might measure the agglomerated state of your sample instead of the primary particle size. 

To address this, we have integrated AI and automation solutions to enable all users to obtain high-quality data without needing extensive experience in measurement best practices. With automation, users do not even need to be in the same room as the instrument when the data is generated. 

This also results in significant time savings. Highly trained and knowledgeable scientists can now focus on other important problems. We often hear from customers that they want their experts to work on more impactful projects, and our solutions facilitate this. 

How has integrating AI and automation solutions influenced standard practices and workflows in industries reliant on particle size analysis?  

I think that is quite a difficult question to answer as particle size analysis is used in so many industries and to date they have each adopted technologies differently. 

Generally, the aim is always to generate better-quality data quickly so users can have more confidence in their results. Ideally, this should be achieved without altering the final outcome, as the underlying physics remains unchanged. 

The materials being tested can be quite complex, often requiring assumptions. With AI, industries can rely less on these assumptions and gain insights into the specifics of their samples. 

Where standardization exists, there is inertia, and it takes time for new technologies and approaches to be adopted. There is increasing use of AI for late-stage processing, where the size result informs decision-making about products and processes. AI in late-stage processing is making it easier for materials scientists to observe correlations between data sets from different measurements and technologies. 

Can you elaborate on how customers benefit from implementing AI and automation in particle size analysis? What improvements have they observed regarding efficiency, accuracy, or other metrics? 

Automation is driving an exponential leap in combined sample throughput and the efficient utilization of scientists’ time. Take the Mastersizer Auto-Lab, for example. This system allows you to queue up the automated analysis of 45 samples. Testing 45 samples manually, with each sample taking approximately 10 minutes, would require 7.5 hours of continuous work—a full working day without any breaks. With the Auto-Lab, setup only takes about 60 minutes, freeing up the rest of your day for other projects and problems. 

A key benefit of automation is reducing human error. AI solutions like those in the Mastersizer 3000+ provide constant monitoring, alerting you early if something is wrong. This prevents wasted time on faulty measurements and data, promoting a ‘right first-time’ approach. 

Could you share some specific examples of AI and automation within the Mastersizer? 

The Mastersizer 3000+ features the most advanced AI and automation solutions of any Mastersizer to date. 

• Data Quality Guidance: This feature functions like a GPS for your particle size analysis journey. It alerts you if you’ve deviated from the optimal path and provides instructions to get back on track, ensuring you achieve high-quality particle size data. 

• SOP Architect: This intelligent method development tool is designed for wet dispersion measurements. It covers all core components of the method development process, providing guidance through a standardized workflow. 

• Adaptive Diffraction: Utilizing machine learning for data assessment, this approach provides more reliable sample results in challenging scenarios, such as bubbles or contaminants in the dispersant. The Mastersizer acquires data at 10kHz, capturing data chunks every tenth of a millisecond. Previously, this volume of data was averaged to produce a single scattering pattern, but now, machine learning allows for the processing of these individual data chunks to determine if they are in a ‘steady state’ or ‘transient state.’ 

• Mastersizer Auto-Lab: This automation feature enables the automated analysis of up to 45 regular samples, including three priority samples, for wet analyses. It handles sample addition, performs size measurements using your chosen method, and cleans the system in preparation for the next analysis. 

• Smart Manager: Serving as an example of automated support, Smart Manager reports basic telemetry information to a cloud-based dashboard. With a smart service agreement, your instrument’s performance is automatically monitored. If performance falls below specifications, you are informed, and remedial action can be taken immediately. Additionally, it provides insights into system usability, helping you understand your ROI and plan your lab processes effectively. 

How do these technologies play a role across your wider portfolio?  

Digital, AI, and automation are areas that Malvern Panalytical has been developing for many decades, and this is evident in the diverse ways they are integrated into our portfolio. 

• Automation and Sample Tracking Technology: Automation, including sample tracking technology, has always been crucial to Malvern Panalytical. Our process automation team builds solutions that can incorporate sample data from multiple technologies. 

• Nanosight Pro: Powered by machine learning, the Nanosight Pro for nano-particle tracking analysis leverages NS Xplorer software to enable automated measurements. This removes subjectivity and provides the highest quality size and concentration data for both light scatter and fluorescence analysis. 

• VIS-NIR Spectrometers and Automated Chemometrics: Our VIS-NIR spectrometers, combined with automated chemometrics in the cloud, provide real-time results for non-specialists, particularly beneficial in industries like agriculture and mining. 

• XRD Analytical Software: Our XRD analytical software is increasingly incorporating data science, offering customers a range of applications for X-ray diffraction and scattering. 

What are the biggest challenges you face when integrating new digital tools and AI capabilities into existing systems? 

Data security is a big challenge, so we prioritize getting this right to achieve the highest standards using the latest protocols. 

The development of AI solutions typically requires significant quantities of data to simulate and model sample behaviors, and the data needs to be relevant. Fortunately, MP’s application expertise is unrivaled, and we speak to our customers often. 

For example, 1 TB of data was required to test Adaptive Diffraction to simulate common problems, which Adaptive Diffraction is designed to solve. There is always the added challenge of educating users about what the digital/AI tool is doing. Some users do not necessarily want a black-box solution; they want to know what is happening. 

Since the implementation of enhanced software tools and automation, what has been the feedback from the scientific community? How has this feedback influenced ongoing development and refinement?  

It is a journey that we make together. Implementing a new way of working can sometimes be a headache for our customers, but those with the vision to drive forward to the next level are grateful that we made that journey with them. They soon forget the pain of change and enjoy the savings and freedoms that a new solution can bring. It is a collaboration. 

Customers always need to verify that a measurement provides true and precise data about their samples. When anything changes, they need to experience the value of that change and verify that the core results remain accurate. 

How do you see digital evolution continuing to shape your products and the industry at large?  

The digital revolution is bringing us closer to customers and their problems, as we are not just selling hardware or software but providing comprehensive solutions. 

Our North Star is the ultimate push-button solution: you put your sample in, tell the instrument what you want to know, and it does everything. It is easy to maintain, works continuously, and utilizes all the tools of AI and automation to achieve this. This approach applies to everything we do for our customers. 

Where can readers find more information?  

• Mastersizer | Laser Diffraction Particle Size Analyzers
• Pixels to Particles Webinar Series
• Introducing Adaptive Diffraction for Mastersizer 3000+

About Paul Senior 

He is interested in the rheological characterization of complex fluids and the physical characterization of materials, including particle size/shape analysis using various techniques (Morphologi 4-ID, NTA, DLS, laser diffraction, differential centrifugal sedimentation) and BET-specific surface area analysis. 

Published on AZoM.com on 30 July 2024