“Formulators can move away from trial-and-error”
Can you explain where predictive sciences are used in the coatings industry?
Sander van Loon: Predictive sciences are used to give parameters to ingredients in order to find compatible ones to formulate with. These parameters work in such a way that ingredients in the same area match with each other. This means, they form something stable and improve efficiency for all types of formulations; either a solution, a dispersion or an emulsion.
If the ingredients are not located in the same region of the parameters, they are not compatible, and you will get problems in stability, film formation and other properties. You can use this when, as an example, you have ten ingredients and want to make sure there is at least some overlap in these values, so that every ingredient is “happy” with the other. This allows to get the best performance of the formulation, with the lowest number of ingredients, at the lowest concentration.
And what kind of parameters do you use to achieve this?
van Loon: Especially for coatings formulations the Hansen Solubility Parameters (HSP) are the best to use. These parameters can be obtained for all relevant ingredients, like fillers, resins, additives etc. It is namely not only restricted to solubility, but also dispersability, or in more general, to find similarity.
Sander van Loon is CEO at VLCI, a service company for formulators and R&D. He will explain this technology at the European Coating Technology Forum | Optimising cost and processes in paint formulation.
The HSP consists of three parameters: dispersion forces, polarity and hydrogen bonding. When practically determined, these center values comes with a sphere, in which other ingredients are compatible or when outside the sphere, are incompatible.
And how do you obtain these parameters?
van Loon: There are two ways for HSP. You can calculate them based on the structure of the ingredients (SMILES). But this is not very precise, and for polymers it is not really representative. The second method is to determine them in a practical manner. Here you have a set of solvents of which you know the HSP, to which the unknown ingredient is added, and then you rate all these samples. Typically, you have to rate between 30 and 40 samples from good to bad solubility, and the rating is then entered into the HSPiP software to do the calculations. In this way you get the HSP and sphere of the unknown ingredient.
This is what you do at your company, so I expect the manufacturers or suppliers of raw materials do not deliver the HSP with their data sheets?
van Loon: No, at the moment, most of them don’t. But I hope this will change soon! We at VLCI are a R&D services company that, among others, determines and implements these parameters for the coatings industry, but also for the personal care, household or polymer industry. We do this for raw material suppliers, so they know more about their own products, and also for formulators, so they can find matching ingredients to improve e.g. the stability of their formulation.
Thus, you can predict the stability of formulations, but you do not get any prediction on the properties of the final product?
van Loon: Yes, stability, but also efficacy, thus properties in many ways. If, for example, you formulate with anti-corrosion pigment and you do not select the best dispersant, your coating will not perform at the maximum. So, matching ingredients via HSP can give you an increase in performance, because you have a better dispersion or solution of your ingredients.
Aside from the HSP you also use the so called HLD-NAC. What is this?
van Loon: Yes, that stands for Hydrophilic Lipophilic Difference – Net Average Curvature that you use for e.g. emulsion polymerisations. You can characterise monomers and, based on the parameters of the monomers, you can find the right surfactant to emulsify them and make your polymer. Basically, it works just like the HSP. You get parameters that you can use to calculate the right ingredients for a specific emulsion.
You mentioned before that you need to test a lot of samples to get the data for just one ingredient. How do you handle this?
van Loon: This is where the high-throughput equipment comes in. In order to practically determine the parameters, you need around 40, sometimes 50 different samples for each ingredient. So, as you know, there are tens of thousands or maybe even hundreds of thousands of ingredients available on the market. To get all these parameters, you need high-throughput. Even in one coating formulation you have around 10 up to 20 different ingredients, which means that you have to create hundreds of samples just for this one case. But once you have the HSP of an ingredient, you don’t have to do that again.
I would imagine that you have a pretty big database on these values by now?
van Loon: There is a database, especially on HSP, but we do all our measurements for customers. That means that most things we do is confidential. Sometimes our customers publish the values, but that is their decision and not ours.
Sometimes we do work for ourselves and then we can publish something, for example, on our website. But, there is a database available with calculated values for HSP. That is an initiative we took together with SpecialChem and Prof. Steven Abbott. We created an ingredient selector on the SpecialChem website, which contains about ten thousands of calculated HSP values. These are not as good as practical determined HSPs, but they work as a guidance and can teach people what they can do.
How complex is it to work with this data? Do you need special skills?
van Loon: A bit complex but do-able. Mostly, it is just a different mindset and takes a bit of time. But once you get an idea of how it works, formulators can start working with it. Also, we give trainings and teach people how to use it.
With HSP, formulators can move away from the trial-and-error approach. A lot of formulators still mix ingredients together, see what happens and then plan further. In some cases, they include design-of-experiments approach to limit the trial-and-error difficulty. But still, it is just shooting in the dark.
And that is the big difference with predictive sciences. Once you have HSP you can do predictions of matching ingredients and limit the number of samples you need to develop your formulation. It is an effort in the beginning, but then the fun begins!
A high-throughput laboratory can be quite expensive. Is this something smaller companies can afford?
van Loon: You can make it as crazy as you want, and it can lead to investments of millions of EUR. But with a smaller budget, let’s say, 100,000 EUR, you can still get a smaller high-throughput that can prepare samples. But in most cases for smaller companies, it makes more sense to outsource to service companies like ours, so you only use it whenever you need. Because for them, it might be that after a week or month of HT screening you would have a lot of data to work with and from then on the HT would be useless for a long period Nevertheless, we also work for large companies, because they want to buy in resources on project basis.
The Interview was conducted by Jan Gesthuizen
If you want to learn more about this technology and other ways to improve modern paint formulation and manufacturing, you should attend the European Coating Technology Forum | Optimising cost and processes in paint formulation in October 2019 in Berlin.