Recent advances in robotics and AI have enhanced many scientific domains with automation benefits such as higher throughput, reduced workload and improved product quality. While common in biology and diagnostics, automation is less explored in colloidal chemistry, particularly in gold nanoparticle preparation. There are still many unresolved issues regarding the reproducibility of the colloidal syntheses and the understanding of the factors in the physical and chemical reaction spaces. One possible explanation for the poor reproducibility is the existence of underestimated, but critical process parameters. Precise control over size and shape is vital for gold nanoparticles’ unique optical properties, especially for the transition between laboratory experiments and commercial products with well-defined properties and quality assurance.
In order to improve the reproducibility of the colloidal synthesis of gold nanoparticles and explore the process parameter space, which may contain previously unknown factors that influence the reproducibility, a synthesis robot was designed and constructed. In contrast to commercially-available liquid handlers, which are expensive and poorly suited for nanoparticle synthesis, it consists of components that are easy to obtain and cost-effective. Many of the parts were rapidly prototyped and produced via 3D-printing. Hardware and software aspects were investigated with a focus on their adaptation and optimization for the successful accomplishing of a reproducible synthesis of gold nanoparticles. Less explored process parameters, such as pipetting speed and exact pipetting position (offset), were also investigated in order to elucidate their role in the formation of particles. It was found that both pipetting speed and offset, which are virtually impossible to control in the manual synthesis, play a role in the resulting average gold cluster size. The size of the clusters has been previously investigated and found to be a determining factor in the size and shape of nanorods, synthesized by seeded growth. This was also confirmed by experiments conducted with differently-sized cluster populations. A critical process parameter was also discovered in the synthesis of gold nanostars, where the short delay between the addition of the growth solution and reducing agent was found to strongly influence the final nanostar size. These results demonstrate a new approach, centered around the precise control of process parameters by robots, for the improvement of the results of colloidal synthesis.