11.11.2025
Jujo Thermal Case Study
Coatings for thermal papers
This case study demonstrates how a leading thermal paper manufacturer partnered with BOMP to optimize the optical density of one of their paper coatings, achieving better results while reducing the need for costly and time-consuming experimental trials. In thermal paper development, performance is typically evaluated by analyzing how optical density changes as a function of applied thermal energy. The relationship between optical density and energy describes how sensitively the coating responds to heat and how efficiently it produces color.
The dataset processed by BOMP consists of dozens of features, including double-digit numbers that represent various pigments used in coatings, rheology modifiers, and coating physical characteristics such as thickness. BOMP's goal is to identify a combination of features that produce better optical densities at increasing discrete energy levels. Analyzing this complex, multi-dimensional data manually is slow, costly, and rarely leads to the best formulation. To overcome this, BOMP analyzed the partner’s dataset to identify and explore optimized coating compositions.
BOMP's machine learning algorithms were trained on the small dataset provided by the partner, which consisted of 33 R&D trials along with their measured optical densities at several energy levels. This created a multi-objective optimization problem where dozens of features needed to be optimized to increase the optical densities across those energy levels. Despite the dataset being relatively small, BOMP suggests and accurately predicts the performance of new, untested coating compositions that meet the design criteria of the partner.
The software operates in about 10 seconds, generating results in a PDF format that displays the optimal combinations of coating formulations with maximized optical densities. The algorithm effectively identifies a set of promising new candidate coating compositions that are predicted to outperform all previous formulations. It intelligently navigates the trade-offs between different ingredients to meet the partner's stringent design criteria, which stem from various sources such as legislation, product standards, and the incompatibility of certain concentrations of inputs. Our partner validated BOMP’s suggestions, which demonstrated performance and metrics closely aligned with BOMP’s predictions.
The algorithm identified the "best" candidate, achieving a significant improvement in the optical density of their coatings. This advancement pushed the performance of the new thermal paper beyond what had previously been accomplished through experimental efforts. Importantly, BOMP did not just identify a single optimal point; it revealed several high-performing combinations that clustered near the optimum, each featuring a different set of coating compositions. This provides the company with valuable options and flexibility, allowing them to select a final formulation based on additional factors such as raw material cost, supply chain stability, and manufacturing compatibility, all while maintaining high product standards. By beginning the next round of physical testing with this curated list of BOMP-validated formulations, the partner can save considerable R&D resources and accelerate their path to market with a superior and safer product.
The dataset processed by BOMP consists of dozens of features, including double-digit numbers that represent various pigments used in coatings, rheology modifiers, and coating physical characteristics such as thickness. BOMP's goal is to identify a combination of features that produce better optical densities at increasing discrete energy levels. Analyzing this complex, multi-dimensional data manually is slow, costly, and rarely leads to the best formulation. To overcome this, BOMP analyzed the partner’s dataset to identify and explore optimized coating compositions.
BOMP's machine learning algorithms were trained on the small dataset provided by the partner, which consisted of 33 R&D trials along with their measured optical densities at several energy levels. This created a multi-objective optimization problem where dozens of features needed to be optimized to increase the optical densities across those energy levels. Despite the dataset being relatively small, BOMP suggests and accurately predicts the performance of new, untested coating compositions that meet the design criteria of the partner.
The software operates in about 10 seconds, generating results in a PDF format that displays the optimal combinations of coating formulations with maximized optical densities. The algorithm effectively identifies a set of promising new candidate coating compositions that are predicted to outperform all previous formulations. It intelligently navigates the trade-offs between different ingredients to meet the partner's stringent design criteria, which stem from various sources such as legislation, product standards, and the incompatibility of certain concentrations of inputs. Our partner validated BOMP’s suggestions, which demonstrated performance and metrics closely aligned with BOMP’s predictions.
The algorithm identified the "best" candidate, achieving a significant improvement in the optical density of their coatings. This advancement pushed the performance of the new thermal paper beyond what had previously been accomplished through experimental efforts. Importantly, BOMP did not just identify a single optimal point; it revealed several high-performing combinations that clustered near the optimum, each featuring a different set of coating compositions. This provides the company with valuable options and flexibility, allowing them to select a final formulation based on additional factors such as raw material cost, supply chain stability, and manufacturing compatibility, all while maintaining high product standards. By beginning the next round of physical testing with this curated list of BOMP-validated formulations, the partner can save considerable R&D resources and accelerate their path to market with a superior and safer product.
